R v Secretary of State for Justice

These appeals concern the statutory provisions governing the eligibility for compensation of persons convicted of a criminal offence where their conviction is subsequently quashed (or they are pardoned) because of the impact of fresh evidence. The provisions are contained in section 133 of the Criminal Justice Act 1988 (“the 1988 Act”) as amended by section 175 of the Anti-social Behaviour, Crime and Policing Act 2014 (“the 2014 Act”). The central issue is whether they are compatible with the presumption of innocence as guaranteed by article 6(2) of the European Convention for the Protection of Human Rights and Fundamental Freedoms (1953) (Cmd 8969) (“the Convention”).

Mr Hallam’s case

Mr Hallam was convicted of murder, conspiracy to commit grievous bodily harm and violent disorder, following a gang fight in which another young man was killed. The case against him at his trial rested on identification evidence provided by two witnesses. The weaknesses in their evidence were such that independent supporting evidence was, in practice, essential. The only support was evidence from a Mr Harrington, denying that he had seen Mr Hallam either on the day of the murder or on the days surrounding it. That evidence was said to suggest that Mr Hallam had concocted a false alibi, since he had stated that he was with Mr Harrington at the time of the murder.

Several years after the trial, the case was referred to the Court of Appeal Criminal Division (“the CACD”) by the Criminal Cases Review Commission on the basis that fresh evidence had been discovered. That evidence included photographs found on Mr Hallam’s mobile phone, showing him with Mr Harrington on the day after the murder. The phone had been seized from Mr Hallam at the time of his arrest but had not been examined. Hallett LJ, giving the judgment of the CACD, observed that this evidence changed the situation dramatically, in that “the evidence relied upon by the prosecution to support the identifying witnesses, namely the evidence as to false alibi”, had been “significantly undermined” ([2012] EWCA Crim 1158, para 75). She went on (para 76):

Summarising the position (in para 77), the court noted that neither identifying witness had been “particularly satisfactory”, with their “various accounts [containing] numerous inconsistencies and contradictions”; and that there was other fresh evidence comprising information provided to the police by a witness named Gary Rees, which had not been disclosed to the defence at the time of the trial, to the effect that another man with the same first name as Mr Hallam was rumoured to be responsible for the murder. The CACD stated (para 77):

The CACD also held (para 79) that, given the terms of the judge’s direction, there was a possibility that the jury might not have realised that it was entitled to treat the evidence of another witness as potentially exculpatory of Hallam. In paras 80 and 83 it stated the conclusion that it drew from all the factors as follows:

Earlier in its judgment, the CACD recorded at para 49 that counsel appearing for Mr Hallam had invited it to state that he was innocent of the offences. The court cited a passage in the judgment of Lord Judge CJ in R (Adams) v Secretary of State for Justice (JUSTICE intervening) [2011] UKSC 18; [2012] 1 AC 48, para 251, as setting out what Hallett LJ described as “the court’s powers in this respect”. The court declined to make such a statement, observing that “we were not satisfied it would be appropriate to use that power on the facts of this case”.

Mr Hallam spent seven years and seven months in prison prior to the quashing of his conviction. He applied for compensation under section 133 as amended. By letter dated 14 August 2014 the Secretary of State refused the application. The letter began by explaining the statutory test:

Two factors were therefore of particular importance: first, that as the CACD had found, the fresh evidence did not establish positively that Mr Hallam was not at the murder scene on the night in question, and secondly, that the CACD had declined to exercise what was described as “its discretionary power” to state that Mr Hallam was factually innocent. The letter concluded:

Mr Nealon was convicted of an attempted rape committed in August 1996. There was identification and description evidence from several witnesses which if accepted placed him in a club where the victim had previously been on the night of the offence, and near the scene of the attack. He denied that he had ever been to the club and gave evidence of an alibi. The victim gave evidence that the man who attacked her “mauled” her, tried to kiss her and put his hand inside her blouse over her bra. He was pulling at her tights and underwear. No DNA examination of her clothing was then carried out.

The case was subsequently referred to the CACD by the Criminal Cases Review Commission on the basis of evidence of DNA found on an examination of her clothing carried out in 2010, nearly 14 years after the offence. A sample taken from the front of her blouse revealed a full male DNA profile from what was probably a saliva stain. It was not from Mr Nealon, but had been deposited by a man who was designated as the “unknown male”. Further probable saliva stains were detected on both cups of her bra. They too had not been deposited by Mr Nealon, but were consistent with the DNA of the unknown male. An examination of her skirt and tights disclosed a complex mixture of DNA, including DNA from an unknown woman, and was inconclusive. Evidence was adduced on behalf of the Crown that the attacker might not have transferred any DNA to the victim’s clothing.

The victim was re-interviewed in connection with the new investigation. She said that she had bought the blouse and bra either on the day of the attack or a day or two before. This was the first time she had worn either garment in public. She had been in a relationship with a male partner at the time, and could not recall any consensual contact with any other man since she bought the blouse and bra. DNA tests excluded the possibility that her partner, any of the officers involved in the investigation, any of the men who arrived at the scene of the attack shortly after it occurred, or any of the scientists involved in the original investigation, was the unknown male. It was argued by the Crown that the DNA might have been deposited on the blouse and bra at the time of their purchase or as a result of re-distribution from other items, and might have nothing to do with the attack, particularly in the light of the victim’s evidence that she had hugged and kissed other men on that date, when she was celebrating her birthday.

The CACD (Fulford LJ, Kenneth Parker J and Sir David Calvert-Smith) concluded that the effect of the fresh evidence was to render the conviction unsafe, and that it should therefore be quashed: [2014] EWCA Crim 574. The central reasoning of the court is found in para 35 of the judgment delivered by Fulford LJ:

Mr Nealon spent 17 years in prison prior to the quashing of his conviction. He applied for compensation under section 133 as amended. By letter dated 12 June 2014 the Secretary of State refused the application. After explaining the statutory test in the same terms as the letter sent to Mr Hallam, the letter continued:

In Mr Nealon’s case, as in Mr Hallam’s, the decision letter focused on the reasoning of the CACD: that it said that the argument that the DNA material might not have been crime-related required serious consideration, that it found that the original jury had been entitled to convict on the basis of the existing identification evidence, and that it said that the fresh evidence did not demolish the prosecution evidence. On that basis, the Secretary of State stated:

Section 133(1) of the 1988 Act provides:

Section 133 was enacted to give effect to the United Kingdom’s international obligations under article 14(6) of the International Covenant on Civil and Political Rights 1966 (“the ICCPR”), ratified by the United Kingdom in 1976. Article 14(6), in its English version, provides:

Section 133(1) restricts compensation to cases where a person’s conviction has been reversed (or he has been pardoned: for the sake of brevity, I will focus from this point onwards on cases where convictions are reversed) “on the ground that a new or newly discovered fact shows beyond reasonable doubt that there has been a miscarriage of justice”. Convictions are not quashed in England and Wales on the ground that there has been a miscarriage of justice, but on the ground that they are unsafe: see further paras 25 et seq below. It was said in Adams, para 36, that the words “on the ground that” must, if they are to make sense, be read as “in circumstances where”, and that the Secretary of State must therefore determine whether a new or newly discovered fact shows beyond reasonable doubt that there has been a miscarriage of justice. In deciding that question, the Secretary of State would have regard to the judgment of the CACD, but ultimately had to form his own conclusion.

The term “miscarriage of justice” was not defined when section 133 was originally enacted. This resulted in a series of cases in which the courts sought to interpret the term, culminating in the decision of this court in Adams delivered on 11 May 2011. In that case, the court adopted four categories of case, of progressively wider scope, as a framework for discussion. They were:

cases where the fresh evidence shows clearly that the defendant is innocent of the crime of which he was convicted;

cases where the fresh evidence so undermines the evidence against the defendant that no conviction could possibly be based upon it;

cases where the fresh evidence renders the conviction unsafe in that, had it been available at the time of the trial, a reasonable jury might or might not have convicted the defendant; and

cases where something has gone seriously wrong in the investigation of the offence or the conduct of the trial, resulting in the conviction of someone who should not have been convicted.

Section 133 was then amended, with effect from 13 March 2014, by section 175 of the 2014 Act, so as to confine the term “miscarriage of justice” to category (1) cases. Section 133(1) remained unaltered: it continued to be necessary for the conviction to be reversed “on the ground that a new or newly discovered fact shows beyond reasonable doubt that there has been a miscarriage of justice”. However, section 175 of the 2014 Act inserted section 133(1ZA) into the 1988 Act, providing a statutory definition of the term “miscarriage of justice”:

As stated already however (para 17 above, and see paras 25 et seq below), the ground on which a conviction is quashed by the CACD is that it is unsafe. Section 133 has therefore to be understood as requiring compensation to be paid only where the Secretary of State determines that the CACD quashed the conviction in circumstances where fresh evidence shows beyond reasonable doubt that the person did not commit the offence.

It was under section 133 as so amended that Mr Hallam’s and Mr Nealon’s applications for compensation were considered and refused by the Secretary of State.

Mr Hallam and Mr Nealon contend that section 133(1ZA) is incompatible with article 6(2) of the Convention, which provides:

On appeal, the Court of Appeal (Lord Dyson MR, Sir Brian Leveson P and Hamblen LJ) considered that it was bound by the decision in Adams to hold that article 6(2) was not applicable to section 133: [2016] EWCA Civ 355; [2017] QB 571. On the other hand, it also considered that the line of Strasbourg jurisprudence including and following the judgment in Allen v United Kingdom(2013) 63 EHRR 10 was so clear and constant that, if not bound by Adams, it would have followed it. The court also agreed with the Divisional Court, for the reasons which it had given, that section 133 was in any event compatible with article 6(2).

The central issue on this appeal can be split into two broad questions:

The first concerns the scope under English law of article 6(2)scheduled to the Human Rights Act 1998: in particular whether and how far it applies at all to decisions on, or the criteria for, the award of compensation under section 133 of the Criminal Justice Act 1988; this question requires us to consider inter alia whether this court should depart from its decision in Adams.

The second question, arising if and so far as article 6(2) is applicable in respect to such decisions or criteria, is whether the definition of “miscarriage of justice” in section 133(1ZA), introduced by section 175 of the Anti-Social Behaviour, Crime and Policing Act 2014 is incompatible with article 6(2).

Before addressing these questions directly, it is appropriate to discuss an underlying question, namely the place of innocence in criminal proceedings.

In English law, as in many other legal systems, it is not the function of criminal proceedings to determine innocence. As Lady Hale stated in Adams, para 116:

It is equally not the function of the CACD on an appeal (or on a reference by the Criminal Cases Review Commission, which is by statute treated as an appeal) to determine whether the appellant did or did not commit the offence. The question for the CACD is whether the conviction is unsafe. Section 2(1) of the Criminal Appeal Act 1968 provides that the CACD shall allow an appeal “if they think that the conviction is unsafe”. The court is then required by section 2(2) to quash the conviction. Section 2(3) provides that an order quashing a conviction shall, except where a retrial is ordered “operate as a direction to the court of trial to enter, instead of the record of conviction, a judgment and verdict of acquittal”. A successful appellant is therefore “in the same position for all purposes as if he had actually been acquitted”: R v Barron [1914] 2 KB 570, 574.

That it is not the function of the CACD to make findings of innocence was emphasised by Lord Phillips in Adams. In his judgment, he expressed agreement with the position as put in the Canadian case of R v Mullins-Johnson (2007) 87 OR (3d) 425, where the Court of Appeal of Ontario said:

Lord Hope and Lord Kerr spoke to similar effect in paras 95 and 172, while acknowledging that the CACD may in practice occasionally “observe that the effect of the material considered in the course of the appeal is demonstrative of innocence”, or make an observation to like effect: see per Lord Kerr, para 172.

Lord Judge, in a dissenting judgment, agreed (para 250) that innocence is “a concept to which the criminal process is not directed”. Hence, he also accepted, the word “‘innocent’ could have no place in section 133”. But he went on in para 251 to say that a CACD was entitled to state that a defendant was innocent and that, if the evidence unmistakeably demonstrated that the appellant was in truth innocent “the terms of the judgment should conscientiously reflect the true reasons for its decision that the conviction should indeed be quashed as ‘unsafe’”.

In relation to Mr Hallam, the CACD spoke of that passage in Lord Judge’s judgment as setting out “the court’s powers”, and decided that it would not be appropriate “to use that power” in Mr Hallam’s case (see para 5 above). The Secretary of State referred to these statements in his own remarks (para 6 above).

It should be made clear that the CACD does not possess any power to make formal findings or declarations of innocence. Nothing in the Lord Chief Justice’s judgment in Adams suggested that it did. It is not the CACD’s role to determine whether the appellant is factually innocent. The question which it determines is whether the conviction is unsafe. When giving its decision on that question, the court will necessarily explain the reasons for its decision. What it is appropriate to say in that regard will depend to a large extent on the circumstances of the case. In practice, it is often necessary to carry out an assessment of the strength of the evidence as a whole, both inculpatory and exculpatory. If the court considers that the evidence plainly exonerates the appellant, then it is entitled to say so when giving its reasons for allowing the appeal. Sometimes the Crown will have accepted that this is so, and in that event the judgment will normally record that stance. In other cases the significance of the fresh evidence is contested, and in that event the court generally confines itself to the issue of safety.

It follows that, although there are some cases in which the court may state in its judgment that the appellant has been exonerated, it is not the purpose of the appeal proceedings to determine whether that is the position, and in the great majority of cases the court does not enter into the fact-finding exercise which would be necessary before such a statement might be made. The absence of any statement that the appellant has been exonerated does not therefore carry any implication concerning the appellant’s innocence.

It is, therefore, highly undesirable that whether the CACD should say that the appellant is innocent of the crime of which he was convicted should become an issue in an appeal, as it became in Mr Hallam’s case. This is not only because the issue does not properly arise. As the Canadian court explained in the case of Mullins-Johnson, it is also important that the significance of acquittals should not be degraded by the introduction of a practice of distinguishing in a criminal context between those who are factually innocent and those who merely benefit from the legal presumption of innocence: a distinction which section 133, in its amended form, can have the understandable but unfortunate effect of encouraging successful appellants to ask the CACD to draw. Cases in which the CACD expresses the view that an appellant was innocent should remain, as Lord Bingham and others have said, very rare. No adverse inference should be drawn from the court’s unwillingness to express such a view. The application of section 133 is for the Secretary of State, not for the CACD quashing the conviction.

Article 6 is headed “right to a fair trial” and article 6(2) reads:

The European Court of Human Rights (the “ECtHR”) has however taken the view that article 6(2) has a continuing relevance after acquittal or discontinuance. In this connection, it recently stated as its starting point these propositions:

Nevertheless, analysing the Strasbourg case law up to 2011 in the course of giving the majority judgment in Serious Organised Crime Agency v Gale[2011] UKSC 49; [2011] 1 WLR 2760 on 26 October 2011, Lord Phillips was inclined to the view that

The ECtHR, without referring to the discussion in Gale, indicated in Allen v United Kingdom on 12 July 2013 that it does not view article 6(2) in so clear cut or limited a sense as Lord Phillips suggested. First, it has developed, as an initial test of the application of article 6(2), the theory of a “link” between, on the one hand, an acquittal or discontinuation of criminal proceedings and, on the other, certain other types of proceedings or claims not involving the pursuit of any criminal charge. A range of cases in which a link has or has not been detected is listed in the ECtHR’s judgment in Allen, para 98. The original concept of a link was, presumably, to set some limit on the expansion of article 6(2) beyond its natural sphere. The ECtHR has however gone on to say that the link may exist either because of the perceived closeness of the subject-matter or simply because of a choice of words used by a court in the other proceedings. So, ultimately, the question whether article 6(2) applies can simply depend on the words used.

Second, where the link is held to exist, the ECtHR has drawn distinctions between (a) claims by a defendant for eg costs or compensation arising out of the termination in his or her favour of the criminal proceedings, and (b) claims by third party victims against a defendant who has been acquitted in criminal proceedings or against whom criminal proceedings have been discontinued. (For the purpose of any such distinction, at least some issues raised by the state would presumably need to be treated as being, in reality, claims by or in the interest of a third party, eg child care proceedings brought by the state.) In the former case, (a), the ECtHR has held that, where there has been “an ‘acquittal on the merits’ in a true sense” (rather than a discontinuance or an outcome sharing features associated by the ECtHR with a discontinuance) any voicing of suspicion of guilt by the public authority against whom such a claim is made constitutes a violation of article 6(2): Sekanina v Austria(1994) 17 EHRR 221 and Allen v United Kingdom, paras 122-123. Even in a case of or similar to discontinuance, it appears, however, from para 128 of the ECtHR’s judgment in Allen, as Lord Reed notes, that nothing must be said in a civil context which calls into question the innocence of the defendant in the criminal context.

The rationale of any distinction between (“true”) acquittals and discontinuance is not easy to understand. If the presumption of innocence is the key, one would have thought it equally applicable in both situations, or possibly even more so in a situation where the state has not felt able to pursue any criminal charges at all and has therefore discontinued. Be that as it may be, the application of any such distinction is itself fraught with difficulties - as is evident by a comparison of Sekanina itself with Allen. In Sekanina, the defendant was acquitted by the jury. The Code of Criminal Procedure required acquittal “where the court finds that … the alleged offence was not made out or that it has not been established that the accused committed the act of which [she] is accused”. In contrast, the statutory condition for awarding costs and compensation in each case depended, in summary, on the absence of suspicion generated by the defendant’s conduct. The Austrian courts made a careful analysis of the circumstances, including the criminal court file, and concluded that this condition was not satisfied. The Austrian Court of Appeal said:

The Austrian courts therefore distinguished between the acquittal and any entitlement to compensation. Nevertheless, the ECtHR said that it was of the opinion that “Austrian legislation and practice nevertheless link the two questions - the criminal responsibility of the accused and the right to compensation - to such a degree that the decision on the latter issue can be regarded as a consequence and, to some extent, the concomitant of the decision on the former” (para 22). Bearing in mind the distinction drawn by the Austrian courts, the suggested consequence and concomitance are both elusive. However, they were only invoked to establish that article 6(2) was engaged, in the sense that it was open to the complainant to assert that it was potentially infringed at all. What was critical is whether it was actually infringed. Here, the ECtHR, after referring to that court’s “comprehensive list of items of evidence against Mr Sekanina” and to the care with which that court had examined the witness statements, and reciting the passage from the Court of Appeal’s judgment, last set out, went on in the critical part of its judgment (para 30):

It appears that the ECtHR not only disagreed with the Austrian Court of Appeal’s analysis of the trial and jury record, but also held it to be illegitimate, in terms of the Convention and in the context of compensation, for the Austrian courts to embark in the first place on any consideration whether suspicions remained in the light of the acquittal. Contrast the ECtHR’s recent judgment in Allen, where the ECtHR upheld the decision of the Secretary of State and of the courts judicially reviewing his decision that it was legitimate to refuse compensation on the ground that the CACD’s setting aside of Ms Allen’s conviction merely established was that the new evidence “might” have led the jury to a different result - meaning that the conviction was unsafe. The jury’s acquittal in Sekanina was evidently analysed as a “true” acquittal or exoneration, whereas the CACD’s was not. But what then would be the position if a criminal judge or court were (as can happen) to acquit a defendant on the basis that the prosecution had not established its case to the requisite criminal standard and/or that the defendant was entitled to the benefit of the doubt? Why should such an outcome at first instance be treated any differently from the outcome before the CACD on appeal in Allen? And, if the two situations are alike, then the potential applicability of Sekanina must, in the light of Allen, be understood as severely limited in scope.

Turning to claims by third party victims against a defendant after acquittal or discontinuance (case (b) referred in para 39 above), the ECtHR’s position is that:

So at first sight claims by third party victims fall outside the scope of the approach the ECtHR has developed for issues arising between the state and a defendant against whom the state has unsuccessfully pursued a criminal charge, leading to acquittal or discontinuance. The qualification, contained in the second quoted sentence, may, according to its text, be read as corresponding with the view taken by Lord Phillips and others including myself in Gale, that is to say that a later civil court must not undermine an acquittal by suggesting that a person ought to have been convicted on the criminal onus: see para 37 above. But, if this is the direction in which the ECtHR is, as one would hope, moving, it is unfortunate that it was accompanied by the citation of problematic authorities discussed further in paras 49-53 below.

Further, the current upshot, in the ECtHR’s own words in Allen, is that:

Although context is all in the law, this degree of vagueness about general principles is indicative of the uncertain and shifting ground onto which the ECtHR’s expansion of the meaning and application of article 6(2) has led.

Like Lord Phillips, with whose judgment in Serious Organised Crime Agency v Gale I concurred, I can however accept that, once criminal proceedings have concluded with acquittal, or, indeed, a discontinuance, no court should in civil or other proceedings express itself in terms which takes issue with the correctness of the criminal acquittal or discontinuance. Such an extension, achieving a degree of harmony with the approach in Strasbourg, seems at least workable and, of course, reflects what one would hope was anyway proper practice. But courts have often - in contexts not involving the pursuit of a criminal charge and using tools and language appropriate to such contexts - to engage with identical facts to those which have led to a criminal acquittal or discontinuance of criminal proceedings. In such circumstances, it is very commonly the case that the standard of proof will differ in the different contexts of criminal and other proceedings. It is, thus, entirely possible that a court may, in a context not involving the pursuit of any criminal charge, find on the balance of probabilities facts which could not be established beyond reasonable doubt in criminal proceedings. The question whether a link exists between the criminal and, say, civil proceedings then appears as a diversion from the real question. The ECtHR may itself be seen to accept that the concept of a link is not critical, because its statement that the words used may themselves create a sufficient link effectively collapses that concept into a consideration of the nature of the words. However, the question remains what nature of words is it permissible to use? The real test is, or should be, whether the court in addressing the civil claim has suggested that the criminal proceedings should have been determined differently. If it has, it has exceeded its role.

If on the other hand, a court has, on the same facts as were in issue in criminal proceedings leading to an acquittal or discontinuance, determined a civil issue (or any issue other than a criminal charge) against the defendant, and has been confined itself to reasoning relevant to that issue, that means, as I see it, that it has applied the law, rather than infringed article 6(2). I do not believe that either the press or the public is wholly ignorant that the criminal standard of proof may on occasions lead to acquittal or discontinuance, in circumstances where the commission of the offence could be established on the balance of probabilities. There have been very well-publicised cases both here and across the Atlantic. There is also a legitimate public interest in such cases being publicly decided and clearly, rather than obscurely, reasoned.

Unfortunately, as it seems to me, the ECtHR has in a number of judgments condemned courts determining a civil issue for accurate descriptions of the elements of an offence constituting a tort simply because such elements also featured in past criminal proceedings. To require a civil court to tergiversate, by using words designed to obscure the fact that the law may find facts proved on a balance of probabilities which were not proved to the standard necessary for criminal conviction, does not assist either the law or the public or the defendant.

Y v Norway(2003) 41 EHRR 87 is an example of a civil court being apparently expected, in the name of article 6(2), to adopt circumlocutions which do no service to transparency. Ringvold v Norway (Application No 34964/97), a judgment issued by the same section in the same constitution on the same day as Orr v Norway (Application No 31283/04), shows to what fine and unsatisfactory distinctions the past case law may lead. Lord Hughes sets out in his para 118 the circumstances in Orr v Norway. The ECtHR’s reasoning there was that:

A reading of the reasoning of the High Court, set out very fully, at para 9 in the report of Orr v Norway, shows the care actually taken by the High Court to explain the difference between the criminal proceedings and the civil claim. I will not set it out in full, but will take it as read and quote only the first and the last two paragraphs, where the High Court said:

I am unable to discern what the Norwegian High Court should, while fulfilling its civil role, have said in order to avoid conferring “criminal law features” on its reasoning and violating article 6(2). The High Court went to great pains to differentiate and so reconcile its treatment of the criminal and civil issues, and the element of violence, although common to both issues, was a critical element in any adjudication of the civil claim, both as to liability and quantum of compensation. The dissenting opinions of Judges Jebens, Nicolaou and Vajić appear unanswerable on these points.

Many of the points I have so far made are also encapsulated in Judge De Gaetano’s separate opinion in the case of Ashendon and Jones v United Kingdom (Application Nos 35730/07 and 4285/08) and his forceful and pragmatic remarks in his separate judgment in Allen v United Kingdom. I note also that in two more recent cases subsequent to Allen, in which the ECtHR recited the principles in Allen and concluded that a sufficient link existed for article 6(2) to be engaged, the ECtHR went on to accept the reasoning and language of the domestic courts as consistent with that article, although it had examined and relied on the same facts as had led to criminal acquittals. In the first case, Vella v Malta (Application No 69122/10) (11 February 2014) following acquittals on charges of theft and receiving, civil issues had arisen from third party claims to the relevant objects. In the second case, Müller v Germany (Application No 54963/08) (27 March 2014), the issue of the applicant’s safety for probationary release had led the court to form a view on facts occurring during a prior period of probation in respect of which the applicant had been charged and acquitted. Both these cases suggest that the ECtHR may be moving towards a limited view of any application of article 6(2) after acquittal, broadly consistent with that suggested by Lord Phillips in Gale: see paras 37 and 47 above. For my part, I would refuse to depart from Adams and Gale, or to follow the case law of the ECtHR, if and insofar as the ECtHR may in the past have gone further - ie further than to preclude reasoning that suggests that the defendant in criminal proceedings leading to an acquittal or discontinuance should have been convicted of the criminal offence with which he was charged. On that basis alone, in my view, these appeals should be dismissed, since nothing in section 133(1ZA) or in the Secretary of State’s rejections of the appellants’ claims to compensation involves any such suggestion.

Assuming that I am wrong about that, and article 6(2) can have some wider application to claims not involving the pursuit of any criminal charge, the question still arises whether section 133(1ZA) is incompatible with article 6(2). The ECtHR in Allen v United Kingdom, para 128, identified the criteria for compensation stated in the original section 133 as being:

The Supreme Court in R (Adams) v Secretary of State for Justice[2011] UKSC 18; [2012] 1 AC 48 identified for domestic purposes the four categories of case which might be suggested to fall within section 133 in its original form, and which I have set out in para 18 above. The Supreme Court held in R (Adams) that section 133, as originally enacted, enabled compensation to be claimed in categories (1) and (2), but not categories (3) and (4).

Allen v United Kingdom concerned what was, in the English domestic terms used in Adams, a category (3) case, ie a case “where the fresh evidence renders the conviction unsafe in that, had it been available at the time of the trial, a reasonable jury might or might not have convicted the defendant”. The ECtHR also treated the case as having some features more akin to discontinuance than to “acquittal on the merits” (see para 39 above). However, that seems to have been so simply because the CACD confined itself to the basic test (whether the conviction was safe) which it was required by statute to apply, and because the Administrative Court and Court of Appeal, in the judicial review proceedings relating to the Secretary of State’s refusal of compensation, proceeded accordingly: see in particular para 134 in Allen, where the ECtHR said:

The ECtHR held in Allen that there had in these circumstances been nothing in the English courts’ treatment of the defendant under section 133 to undermine her acquittal or demonstrate a lack of respect for the presumption of innocence which she enjoyed, and so no violation.

The ECtHR approached Allen on the basis of the language used by the English courts, rather than an examination of the meaning of section 133. Thus, it said (para 129), that:

Differing views had been expressed in R (Mullen) v Secretary of State for the Home Department[2004] UKHL 18; [2005] 1 AC 1 as to whether section 133 as originally enacted confined the right to compensation to category (1) cases, ie “cases where the fresh evidence shows clearly that the defendant is innocent of the crime of which he was convicted”. That was Lord Steyn’s view, with which Lord Bingham did not associate himself.

The ECtHR’s focus in Allen on the language used by the English courts was possible because it was not suggested in Allen that Ms Allen’s case fell into any category other than category (3): see further paras 67-69 below. The ECtHR did however give a strong clue as to its thinking on the potential consequences under article 6(2) of Lord Steyn’s construction of section 133, had the English courts relied on and applied that, when in para 133 it said:

The new section 133(1ZA) confines compensation to circumstances where a conviction is reversed by the CACD (or a pardon granted) “on the ground that a new or newly discovered fact shows beyond reasonable doubt that there has been a miscarriage of justice” in the sense that it “shows beyond reasonable doubt that the defendant did not commit the offence”. It therefore confines compensation to cases within category (1), matching Lord Steyn’s view of its original meaning. Does this mean that we should declare it to be incompatible with article 6(2)? I readily acknowledge that this might at first sight appear to be the implication of the ECtHR’s thinking in the passage cited above from para 133 of the ECtHR’s judgment in Allen. But the point has never been directly before or decided by the ECtHR, and I am far from confident that its implications have been worked through in a manner which makes it acceptable, or that the ECtHR would conclude that section 133(1ZA) is incompatible if the question were argued out before it.

The first matter that I would address is the clear understanding of the drafters of A3P7, which (although the United Kingdom has not ratified that Protocol) is clearly the origin of section 133: see para 16 above. That understanding appears in the Explanatory Memorandum which was prepared along with the draft Protocol by the Steering Committee for Human Rights, which submitted both documents together to the Council of Ministers on 22 November 1984, the date on which the Protocol was adopted. The Explanatory Memorandum makes clear that A3P7 contemplated just such a provision as now exists under English law in section 133(1ZA). It says:

The ECtHR in Allen addressed this by saying in para 133:

As para 71 sets out, the full text of the Explanatory Memorandum was to the effect that it

In what follows, however, I shall approach the construction of section 133 independently of the Explanatory Memorandum. It might have been thought that, both in its original and in its current form, section 133 (as also A3P7) makes any right to compensation entirely dependent on the ground on which the criminal court (here the CACD) reverses the conviction (or on which a pardon is granted). That would, if correct, have had two consequences. First, it would have marked another distinction from Sekanina, where the award of compensation depended on its face on an independent evaluation of the position by a civil court. Since compensation would then simply have depended on how the criminal court expressed itself, the principle that neither the state nor a later court dealing with a civil claim should say anything different from the criminal court acquitting the defendant could not apply at all. Second, it would have meant that the present appellants had no claim, since a reading of the grounds on which the CACD allowed their appeals indicates that in each case it did so simply because the newly discovered facts made their convictions unsafe. In other words, the CACD’s actual decision was, as in Allen, simply that their cases fell domestically within category (3).

I am not, however, prepared to accept such a construction of section 133 as correct. First, I note that section 133(3) provides that:

It is clear from Allen v United Kingdom that there is nothing wrong with a criminal court, when setting aside a conviction, confining itself (in accordance with its role explained in paras 26 to 34 above) to indicating that “the new evidence, when taken with the evidence given at trial, ‘created the possibility’ that a jury ‘might properly acquit’ the defendant”; or explaining that “the evidence which was now available ‘might, if it had been heard by the jury, have led to a different result’”; or expressing itself in terms which “did ‘not begin to carry the implication’ that there was no case for the applicant to answer”; or indicating that “there was ‘no basis for saying’ on the new evidence that there was no case to go to a jury”: see paras 131-132 in Allen.

All these are ways of expressing a conclusion that a case falls within category (3). They amount to saying that some ground for suspicion remains. Yet it is clear from Allen that they are acceptable and that Sekanina does not have contrary effect. A central plank of the ECtHR’s judgment in Allen is that there is nothing wrong with a refusal of compensation on the ground that the case falls within category (3). That is, as I read both the CACD’s judgments, also the ground on which the CACD allowed both the present appellants’ appeals in the criminal proceedings, as well as the ground on which the Secretary of State disallowed their claims for compensation.

It follows, as the other side of the coin from what I have already said, that the right to compensation can legitimately be expressed to depend upon whether (adopting the terminology in Adams) the conviction was set aside on a ground falling within category (1) or (2). Logically, a defendant wishing not merely to have a conviction set aside, but also wishing to recover compensation, must, unless the case is one of the rare cases (see paras 32 to 34 above) in which the CACD expresses its judgment setting aside the conviction in terms going further than a conclusion that the conviction is unsafe, persuade the Secretary of State to go further. In the rare case where the CACD does express itself in terms stating that the defendant is innocent, that will in practice be conclusive. The Secretary of State could not realistically go behind such a statement. But in other cases, where the CACD has merely determined that the conviction is unsafe, it must be open to the state to resist a defendant’s suggestion that the case falls within a different category that would entitle him to compensation, and for the Secretary of State to reach a conclusion on that basis. Otherwise, as soon as a defendant argues that the Secretary of State should go further than the CACD has gone and should view the circumstances as falling within a category for which the legislature has prescribed compensation, the state would have to accept this, and concede liability to pay compensation. This situation did not of course arise in Allen, because there was no attempt there by Ms Allen to bring her circumstances into any category other than that of category (3) within which the CACD had seen it as falling.

A defendant seeking compensation after the setting aside of his or her conviction by the CACD may therefore be required to show that the circumstances were not merely such that his conviction was unsafe. Using the terminology in Adams, the circumstances must be shown to fall within a higher category, which must, necessarily (and using the terminology in Adams), be either category (1) or category (2), or, since the enactment of section 133(1ZA), category (1) alone. Is there, in terms of compliance with the Convention, any sensible distinction between categories (1) and (2)? Category (1) is no more than a subset of category (2). If it is legitimate for the state to require a defendant to show at least that his or her case falls within category (2), on what basis could it be illegitimate for the state to require a defendant to show that it falls within category (1)? Putting the matter the other way around, the ECtHR has in para 133 in Allen implied that there would be an objection to requiring a defendant to show that the case fell within category (1). But it has not (at least in terms) addressed category (2). It may be that the ECtHR’s passing reference in para 133 to the inappropriateness of Lord Steyn’s test should be understood as embracing both categories (1) and (2). If so, then, as the preceding paragraph of this judgment shows, the effect would be largely to undermine the outcome of Allen itself. All that an applicant for compensation would need to do was assert this his or her claim fell into a higher category than category (3), and the state would be precluded from asserting the contrary, because to do so would be to infringe the “presumption of innocence”.

A way out of this impasse might exist if a sensible distinction could in the context of the Convention be drawn between categories (1) and (2). The legislation, or the language of the courts, could then be amended to speak not of proof of innocence, but of proof that the new or newly discovered fact so undermined the case against the applicant that no conviction could possibly be based on it. But could reference to a case as falling within category (2) sensibly be distinguished from whatever may be thought to be the ambit of the ECtHR’s implied objection to language bringing a case within category (1)? If, to use the ECtHR’s further words in Allen, para 136, it demonstrates “a lack of respect for the presumption of innocence which [a defendant] enjoys in respect of the criminal charge … of which she has been acquitted” to refuse compensation on the ground that the defendant has not shown innocence, it would presumably also demonstrate a lack of respect for the presumption of innocence to refuse it on the ground that the defendant had not shown that she was not only acquitted, but also that there was no evidence upon the basis of which she could possibly have been convicted. The two situations are distinct as a matter of domestic criminal law, and the legislature has distinguished between them for the purposes of compensation. But to distinguish between them in terms of the Convention and in relation to the question of infringement of the presumption of innocence, would seem to do no more than add another fine and unconvincing distinction, in an area where the application of the Convention already appears too full of unsatisfactory and unsatisfying distinctions and uncertainties.

I cannot therefore see any logical basis on which section 133(1ZA) can or should be seen as incompatible in terms of article 6(2) of the Convention. As to the relationship between this court and the European Court of Human Rights’ jurisprudence, I am of course very conscious of what has been said by Lord Neuberger and myself in the passages cited by Lord Reed in his para 172. Like Lord Wilson, I would, however, draw attention to the further words of Lord Hughes and myself in R (Haney, Kaiyam and Massey) v Secretary of State for Justice[2014] UKSC 66; [2015] AC 1344, para 21, where we said that:

Speaking for myself, I cannot regard the current state of European Court of Human Rights’ case law as coherent or settled on the points critical to this appeal. The second point has never been directly addressed; it is at most addressed indirectly by a passing dictum, uttered in a context in Allen where no detailed analysis was necessary because the point did not directly arise. I do not share Lord Wilson’s view, in para 94(c) of his judgment, that it is over-optimistic to suppose that the ECtHR will not think again in relation to article 6(2), generally or, at the least and critically, in relation to its dictum regarding Lord Steyn’s approach quoted in para 49 above. But, however that may be, I question whether the area of law currently under discussion is one where uniformity of approach is critical, even if the precise implications of the ECtHR case law were clear.

In summary, I am, for the reasons given, persuaded that it would be inappropriate to introduce into English law an interpretation of article 6(2) going beyond that identified by Lord Phillips, as set out in paras 37 and 47 above. But, in any event and even if article 6(2) does have a wider application in respect of claims not involving any criminal charge, I am not persuaded that section 133(1ZA) can or should be regarded as incompatible with article 6(2).

For all these reasons a declaration of incompatibility is in my opinion inappropriate.

In general, where it is clear that the European Court of Human Rights would find that the United Kingdom has violated the Convention in respect of an individual, it is wise for this court also to find that his rights have been breached. The object of the Human Rights Act 1998 was to “bring rights home” so that people whose rights had been violated would no longer have to go to the Strasbourg court to have them vindicated. I was initially disposed to think, for the reasons explained by Lord Reed, that the Strasbourg court would indeed find a violation in this case. However, I am persuaded that this is not as clear as once I thought it was, for several reasons.

There are, of course, all the objections in principle to applying the presumption of innocence to any proceedings taking place after the criminal charge has been determined, either by acquittal or discontinuance, so eloquently voiced by Lord Wilson and Lord Hughes. But it is surely too late in the day for the Strasbourg court to revisit that whole question. Furthermore, as Lord Reed has demonstrated, all the arguments deployed by the majority in Adams in holding that article 6(2) was simply not engaged in section 133 cases have been comprehensively rejected by the Strasbourg court. I would therefore agree with him that article 6(2) is engaged in this case.

However, it does not follow that the Strasbourg court would automatically find that it has been breached in this case. As Lord Mance explains (para 39), the Strasbourg court has drawn a distinction between (a) claims by a defendant for such things as costs or compensation arising out of the termination of a criminal case against him in his favour, either by acquittal or discontinuance, and (b) civil claims by or on behalf of third party victims against a former defendant in criminal proceedings which have been determined in his favour. In category (b) cases, where the parties are different, the standard of proof is different, the admissible evidence may also be different, and liability is not dependent upon criminal proceedings having been brought at all, the Strasbourg court has clearly accepted that the civil claim may be determined differently from the criminal proceedings without violating article 6(2). The important thing is the language adopted by the court when deciding the civil claim, as illustrated in the contrasting decisions in Ringvold v Norway (Application No 34964/97), and Y v Norway (2003) 41 EHRR 87. Lord Mance suggests that “the real test is, or should be, whether the court in addressing the civil claim has suggested that the criminal proceedings should have been determined differently” (para 47). I agree, and I share his regret that, in Orr v Norway (Application No 31283/04), judgment of 15 May 2008, the Chamber, by a narrow majority, appear to have asked more of the civil court than this. While accepting that an acquittal in criminal proceedings is no bar to a civil claim for compensation based on the same facts, they appear to have demanded that the court hearing the civil claim phrase its decisions in less than fully transparent language. This is contrary to the rule of law: courts must always be able to explain their decisions fully, clearly and honestly. The one thing they must avoid is suggesting, in civil proceedings, that the defendant should have been convicted of the criminal offence. But I take comfort from the fact that this was the decision of a Chamber of the court, and by the narrow margin of four to three.

This is not a category (b) case, but Lord Mance detects signs that the Strasbourg court might also be prepared, despite the breadth of its language in Allen v United Kingdom (2013) 63 EHRR 10, to adopt an approach to category (a) cases which in practice requires merely that the court determining the defendant’s claim for costs or compensation refrain from any suggestion that he should have been convicted of the offence. There is enough in the evolution of the court’s jurisprudence to suggest that, for the most part and with some limited exceptions, that is in fact what they are doing.

If that were indeed to be the approach of the Strasbourg court to these cases, it might still be that the insistence on showing beyond reasonable doubt that the claimant did not commit the crime in section 133(1ZA) of the Criminal Justice Act 1988 will lead to a violation of article 6(2) in some cases where compensation is denied. But I am not convinced that it would always do so. An indication is the “strong clue” in para 133 of Allen in relation to Lord Steyn’s test (later adopted in section 133(1ZA)), quoted by Lord Mance at para 63. But, as he points out, the court was not addressing such a case in Allen, which was acknowledged to be a case in Adams category (3), where the conviction was quashed because it was unsafe in the sense that the fresh evidence meant that a jury might or might not have convicted. Provided that this is explained without suggesting that the defendant should have been convicted, there is no breach of article 6(2).

The cases before us are also cases, like Allen, in which the fresh evidence rendered the conviction unsafe, in the sense that, had it been available at trial, a reasonable jury might or might not have convicted the defendant. The Grand Chamber found no violation in the case of Allen. In my view, the issue of incompatibility would be better addressed in a case which fell clearly within category (2), where it might be difficult to explain the difference between that and a category (1) case without casting doubt on the acquittal. But if it be right that the true question is whether the Secretary of State, or a court in judicial review proceedings, has suggested that the defendant ought to have been convicted, then it does not seem impossible to explain a refusal of compensation without doing this.

Furthermore, where a particular statutory provision may or may not lead to a violation, it is not appropriate, in my view, to make a declaration of incompatibility in proceedings brought by an individual in respect of whom the Strasbourg court is unlikely to find a violation, as I believe these to be. I should add that my view of the appropriateness of making a declaration of incompatibility in this case has nothing to do with my view of the merits of the amendment to section 133.

My view is that the present appeals place the court in a deeply uncomfortable position.

We afford profound respect to the decisions of the ECtHR and recognise its unparallelled achievements in raising the standards according to which member states of the Council of Europe, undoubtedly including the UK, must treat their people.

I am, however, persuaded that, in its rulings upon the extent of the operation of article 6(2) of the Convention, the ECtHR has, step by step, allowed its analysis to be swept into hopeless and probably irretrievable confusion. An analogy is to a boat which, once severed from its moorings, floats out to sea and is tossed helplessly this way and that.

In what follows I seek to summarise my reasons for this grave conclusion:

I turn to this court’s duty under section 2(1)(a) of the Human Rights Act 1998 [the 1998 Act] to “take into account” any relevant judgment of the ECtHR. Inevitably there have been a number of observations in this court, and in the appellate committee which preceded it, that the duty to take account of such a judgment should almost always lead our domestic courts to adopt it. Particularly in the early years of the life of the 1998 Act, the UK courts were strikingly loyal to the judgments of the ECtHR notwithstanding the open texture of section 2(1)(a): see Krisch, The Open Architecture of Human Rights Law [2008] 71 MLR 183, 203.

In R (Alconbury Developments Ltd) v Secretary of State for the Environment, Transport and the Regions[2001] UKHL 23, [2003] 2 AC 295, the appellate committee perceived no need to confront, as problematic, the jurisprudence of the ECtHR in relation to the relevant article of the Convention, which was article 6(1). It applied it without apparent difficulty. But Lord Slynn of Hadley observed at para 26:

On other occasions this court has expressed the proper approach to the jurisprudence of the ECtHR in different terms. In para 173 below Lord Reed quotes in particular from para 48 of the judgment of the court delivered by Lord Neuberger of Abbotsbury MR in the Manchester City Council case and from para 27 of the judgment of Lord Mance in the Chester case. In my view however the weight to be given to both quotations was correctly described by Lord Mance and Lord Hughes in their joint judgment in the Kaiyam case, cited by Lord Mance in para 72 above, as follows:

The context of the present appeals, to which the nature of this court’s duty under section 2 is therefore specific, is a line of jurisprudence in the ECtHR which - in my respectful view - is not just wrong but incoherent. Our courts have not, to the best of my knowledge, previously been called upon to address a context of that sort.

In In re McKerr[2004] UKHL 12, [2004] 1 WLR 807, Lord Hoffmann said:

I reluctantly agree with Lord Reed, for the reasons he gives in paras 183 to 191 below, that, if article 6(2) has the meaning ascribed to it by the ECtHR, in particular in the Allen case, section 133(1ZA) of the 1988 Act is incompatible with it. It follows that I am at present not persuaded by the ingenious suggestions to the contrary made by Lord Mance in paras 61 to 71 above and by Lord Hughes in paras 128 and 129 below.

But I have come to the conclusion that this court should not adopt the meaning ascribed to article 6(2) by the ECtHR. I have been driven to the view that it should today dismiss the appeals.

(a) I hold in high professional regard our fellow judges in the ECtHR.

Narrowly stated, the question raised by the present appeals is whether the new section 133(1ZA) Criminal Justice Act 1988 is incompatible with article 6(2) of the European Convention on Human Rights (the presumption of innocence). That question can, however, only be answered in the context of the true scope of the presumption of innocence, which arises also in many other legal scenarios. This is a matter with which the Strasbourg court has been obliged to grapple over the past 30 years. The presumption of innocence is also central to the approach of all three UK jurisdictions to the criminal law, as it is to a great many other legal systems.

Article 6(2) provides:

There is no doubt that this governs the investigation of, and the trial of, criminal charges. Centrally, it means that the burden of proof in a criminal trial lies upon the prosecution to show that the accused is guilty rather than upon the accused to show that he is not. In most if not all European systems that rule is associated with a requirement that proof of a criminal offence must achieve a high standard before a defendant can be convicted - usually described as proof beyond reasonable doubt.

There is no occasion to examine this central core of article 6(2), which is not in issue in the present case. What is in issue arises, not for the first time, not from the plain meaning of a Convention right, but from the manner in which it has been extended, by way of judicial gloss, beyond the investigation and trial of criminal offences to legal situations where no charge remains pending and no trial is in contemplation. This gloss is referred to in the Strasbourg jurisprudence as the “second aspect” of article 6(2). Like other judicial glosses, this one has developed piecemeal. That is often the result of iterative consideration of individual cases, but that process needs also to provide the opportunity to stand back and to examine the logical and jurisprudential basis for the steps which have been taken.

It appears from the Grand Chamber’s recent formulation of this “second aspect” of article 6(2) in Allen v United Kingdom(2013) 63 EHRR 10, that it has the features here set out.

The central reality which has to be addressed by any legal test for the scope of article 6(2) is that the same factual issues which have to be decided in a criminal trial or investigation in order to reach a verdict of guilty or not guilty, or a decision as to prosecution, may also have to be decided for other legal purposes. Those other legal purposes may well involve the person who was the accused in the criminal trial or investigation. The other legal purposes may be sequential to the criminal trial or investigation (for example an application for costs) or they may be separately constituted (for example professional disciplinary proceedings against the accused or child care proceedings concerning his children). Some legal systems may adjudicate upon those other legal purposes in combined criminal and civil proceedings, by permitting the complainant in the criminal trial also to make a claim for compensation as a civil party; other systems may adjudicate upon them separately.

It is an axiomatic feature of some legal systems that the law recognises that the enhanced standard of proof required to justify conviction of a criminal offence and punishment by the state does not apply except to the verdict of guilty or not guilty. Elsewhere, the standard of proof is lower, often on the balance of probabilities. There is a well-understood principled basis for this difference. In criminal proceedings the chief object is the punishment of the guilty. Where the state seeks against an individual a conviction and punishment the individual is entitled to the benefit of a reasonable doubt: thus acquittal may well be in dubio pro reo, rather than involve a positive finding that the act alleged was not performed. That this should be so is a proper reflection of the gravity of a criminal conviction. Where, on the other hand, the issue arises between citizens of equal standing before the law, the object is not punishment but compensation or vindication and it unfairly constrains the rights of the claimant if he can succeed only if all reasonable doubt is eliminated. Likewise, the object of professional disciplinary proceedings differs from that of criminal proceedings; where the objective is the protection of the public from unsuitable practitioners it is legitimate and principled to give that protection where it is demonstrated to be more likely than not that it ought to be provided. A fortiori, where the object of proceedings is the protection of the vulnerable, typically but not only children, the criterion for decision is the best interests of the vulnerable and to limit protective orders to cases where maltreatment has been proved beyond reasonable doubt would be inconsistent with that ruling principle.

The three legal systems operating in the United Kingdom all depend upon this marked and principled difference between proof beyond reasonable doubt as a minimum for conviction and punishment and proof on the balance of probabilities in most other areas of adjudication. So do some other European systems, for example Norway: see Reeves v Norway (Application No 4248/02), 8 July 2004 and Orr v Norway (Application No 31283/04), 1 December 2008. The distinction between the two standards of proof may not be as clearly acknowledged in some other European systems (see for example the discussion by Kaplow (2012) 121 Yale Law Journal 738) but it is of course well understood and explained by the Strasbourg court in, for example, Orr v Norway at para 26.

Once the difference in standard of proof is recognised, it is plain that those proceedings to which the civil standard apply simply cannot be governed also by the criminal standard, nor thus by the verdict of the criminal court, even if the same factual issues arise, and even if the evidence is the same. Discussions about the scope of article 6(2) must necessarily accommodate this fact.

The summary of the Strasbourg jurisprudence helpfully set out in Allen v United Kingdom (see para 99 above) might suggest an established and consistent two-stage approach. First, that the concept of link is the test for the applicability of article 6(2) to proceedings. Second, that whilst there is no single test for whether, if applicable, that article is infringed, the critical question is whether the unconvicted accused is treated by a court or public body as if guilty and the language used will generally be of critical importance. The history shows that this is not quite how the cases have proceeded. It demonstrates that the court has grappled frequently with the inevitable tension between the desire to protect an unconvicted accused from having his acquittal undermined and the reality that the outcome of the criminal trial cannot govern all adjudication upon the same factual issues.

The concept of link was not articulated in the early cases, and certainly not in the detailed terms now enunciated in Allen. That is perhaps because the early cases concerned claims for costs and/or compensation for detention on remand in systems such as Germany and Austria where those claims fell to be determined by the criminal courts, indeed sometimes by the same constitution which returned a verdict of guilty or not guilty. Minelli v Switzerland (1983) (Application No 8660/79) is an example, where the criminal court, in acquitting the accused, on the grounds of expiry of the relevant limitation period, also in the same judgment apportioned costs as between the private prosecutor and the accused. It took the view that both were partially at fault. As to the accused, it expressed the view that although he had a limited justified complaint against the prosecutor, the terms in which he had expressed it would have left him in all probability guilty of the criminal libel alleged, but for the limitation period. It was enough for the Strasbourg court to say that at the time when these conclusions were expressed the accused was still “charged with a criminal offence” (para 32).

The next stage was a trio of German cases, all decided on the same day in 1987: Lűtz v Germany (1988) 10 EHRR 182, Nölkenbockhoff v Germany(1988) 10 EHRR 163and Englert v Germany (1991) 13 EHRR 392. All were cases in which the criminal proceedings had been discontinued, in Lűtz because a limitation period had expired, in Nölkenbockhoff because the accused had died whilst appeal against conviction was pending, and in Englert because the much-convicted accused was not likely to receive a significant addition to a sentence he was already serving. In each case, the local court, exercising a discretion plainly given to it by domestic legal rules, had declined either in whole or in part to make orders for costs and/or compensation for detention on remand. In each case the court had ruled in making that decision either that the accused would “almost certainly” have been convicted but for the technical bar which led to discontinuance (Lűtz and Nölkenbockhoff) or was “clearly more likely” to have been convicted and had brought suspicion on himself (Englert). As in Minelli, the Strasbourg court referred to the fact that the decision on discontinuance accompanied that on costs etc, which it described as a consequence and necessary concomitant of the former (eg Lűtz para 56). It then held as to infringement that such costs or compensation issues “might raise an issue under article 6(2) if supportive reasoning, which cannot be dissociated from the operative provisions, amounts in substance to a determination of the accused’s guilt”: Lűtz para 60, repeated in the other cases. In all these cases, nevertheless, the court held that the language used had not infringed article 6(2) because it amounted to no more than voicing outstanding suspicion that the accused had committed the offences, rather than amounting to a finding of guilt (Lűtz para 62, echoed in the other cases). That would appear to have been a plain recognition of the fact that to say of an accused that he might have committed the offence, or even that he probably did, is not to undermine his acquittal, and does not amount to attributing guilt to him. That is even more clearly the case in systems such as the English where an acquittal means no more than that guilt has not been proved to the high criminal standard, may well leave open the possibility that the accused might have committed the act, but establishes once and for all that he is unconvicted and cannot be punished.

The origins of the concept of link, as adumbrated in due course many years later in Allen, may be the two cases of Sekanina v Austria (1993) 17 EHRR 221 and Rushiti v Austria (2001) 33 EHRR 56.Both concerned applications by accused who had been acquitted at trial for compensation for detention on remand. The domestic law provided that compensation was payable if the accused was acquitted “and the suspicion that he committed the offence is dispelled”. The local courts had held that despite acquittal, suspicion had not been dispelled; there had been a strong case, but the evidence had not been enough to convict. The Strasbourg court held both that article 6(2) applied and that it had been infringed. It held that although the court determining the compensation issue had done so some months after the acquittal, nevertheless “Austrian legislation and practice link the two questions … to such a degree that the decision on the latter issue can be regarded as a consequence, and to some extent the concomitant of the decision on the former.” (Sekanina para 22, repeated in Rushiti). Although, as has been seen, the word “concomitant” had also appeared in the three German cases, there is nowhere any analysis of why it is appropriate. It may well be that the decision upon those issues could properly be described as sequential to the verdict, in the sense that a verdict of acquittal was a sine qua non of it, but it does not follow that it was a concomitant or had to run with the verdict; on the contrary the fact that the legal test was different surely meant that it did not run with the verdict. To say that article 6(2) made it run with the verdict would be to assume what was sought to be shown.

Sekanina and Rushiti also broke new ground on the question of infringement. At paras 27 and 30 of Sekanina the court distinguished the three German cases, where the language used had been rather more forthright than in the instant case; it had spoken of it being nearly certain that the accused would have been convicted, rather than of suspicion not having been dispelled. The court held that the approach of the German cases to what had there been regarded as a recording of suspicion only applied to discontinuance cases and not to acquittals. At para 30 it said this:

Since then the link proposition has indeed been oft repeated, generally in identical language, up to and including in Allen. But at no stage has the court gone back to principle to examine what the true scope of article 6(2) is, given the differing legal contexts in which the same facts may be adjudicated upon according to different tests and subject to different standards of proof. Nor are the cases consistent. In Moullet v France (Application No 27521/04), 13 September 2007, the court held that article 6(2) did not apply to disciplinary proceedings taken against a public official for bribery, although he had been acquitted (on limitation grounds) of that offence by the criminal court. It also held that in any event there was no breach, although the act of bribery in question in each set of proceedings was identical and his dismissal was explicitly grounded upon it. But in Vanjak vCroatia (Application No 29889/04), 13 January 2010 and Šikić v Croatia (Application No 9143/08), 15 July 2010 disciplinary proceedings against policemen were held to be linked to criminal prosecutions which had been abandoned, so that article 6(2) did apply; there is no sign that Moullet was referred to.

The test for applicability appears, if anything, to have widened, since as the court recorded in Allen at para 102, these two cases of Vanjak and Šikić contain the opinion that following discontinuation of criminal proceedings (as well as following acquittal) the presumption of innocence requires that the lack of a person’s criminal conviction be preserved “in any other proceedings of whatever nature”. These very wide words are not further reasoned, nor is the apparent departure from the German and Austrian cases explained, and in neither case was the statement necessary to the decision since the applicant failed in both cases on the grounds that the constituent elements of the disciplinary or employment complaints differed from the legal ingredients of the criminal charges which had been discontinued.

There is no doubt that there are relatively recent decisions in the Strasbourg court which, if their approach to article 6(2) were applied to the present case, would result in a finding that section 133(1ZA) is incompatible with that provision. An example is Capeau v Belgium(2008) 46 EHRR 25. The accused had been investigated for suspected arson but discharged by the court on the grounds that there was insufficient evidence to commit him for trial. He claimed compensation for pre-trial detention on remand. Under the local law, there was a right to such compensation either if the accused was “exculpated” by the criminal court (which he had not been) or if he established his innocence. The local law illustrates the variation across Europe of entitlement to compensation for pre-trial detention on remand. The Belgian court refused the application for compensation on the grounds that the accused had not established his innocence. The Strasbourg court held that to refuse compensation could not by itself amount to a breach of article 6(2) but that the requirement that the accused prove his innocence did so. It concluded that this provision “allowed doubt to attach itself to the correctness of the court’s decision”. But that last statement is surely not accurate. To say that someone has not proved himself eligible within the rules for compensation for detention is not, in any meaningful sense, the same as doubting the correctness of a decision that there was insufficient evidence to commit him for trial. Like some other general statements appearing in the article 6(2) cases, it demonstrates a reluctance to address the meaning of acquittal. It may be that in some legal systems an acquittal, and a fortiori a decision not to commit for trial, is a finding of positive exoneration, but in most it is not. It is especially unlikely to be so where the verdict is that of a jury which returns a binary verdict but does not deliver a judgment making individual findings of fact.

In Műller v Germany (Application No 54963/08), 27 March 2014 the claimant was a life sentence prisoner after shooting his wife. He had sought early conditional release. He had recently been charged with assaulting and injuring another woman with an electric truncheon whilst on home leave, but the local criminal court had dismissed the charge without giving reasons. The execution of sentence court, fulfilling a role similar to that of the Parole Board in England, refused his application for conditional release on the grounds that he remained a serious risk to the public and particularly to women. He had become obsessed both with his wife and with the recently-injured woman, and injuries had followed disappointment. The execution of sentence court had additional psychiatric evidence, but it specifically addressed the recent allegation of assault against the second woman, and explicitly disagreed with the criminal court, which it held had not adequately examined the evidence against the accused. It said in terms that “the criminal offence which the applicant had committed” towards the recent complainant woman, demonstrated the risk of violence. The Strasbourg court faithfully applied the general statement made in Allen and found in consequence that there was a sufficient link between the acquittal and the decision on conditional release. But it held that there was no breach of article 6(2): the execution of sentence court had not, it held, stated that the accused was guilty of a fresh offence. Rather it had based its conclusion on the prognosis of risk for the future. It must be said that this obviously correct outcome was reached in the teeth of the words used by the execution of sentence court. Certainly it had based its conclusion, correctly, on the prognosis of risk for the future, but it had arrived at that prognosis in large part because it expressed itself satisfied that the accused had committed the recent offence of which he had been acquitted. A set of principles which requires such specialised reasoning in order to justify an obviously correct conclusion that the assessment of risk involved no breach of article 6(2) puts in issue the basis of the principles.

This case is a remarkable illustration of the consequences of the wide propositions which have developed in the court’s jurisprudence as to article 6(2). It might be thought axiomatic that the assessment of the future risk posed by a convicted murderer whose conditional release is under consideration ought to be informed by all relevant information, and that to exclude material because it reveals the possibility of a criminal offence simply because there is not sufficient evidence to prove it beyond reasonable doubt is to court danger to the public. The much more logical basis for the outcome of the case is surely that a presumption of innocence has no place in such risk assessment. Article 6(2) has no application, for conviction and punishment are not in question. This is so even if on a different legal test and applying a different standard of proof, a conclusion is reached which includes a finding that acts amounting to an offence are relevant to that assessment. The accused in this case was not treated by the legal system as convicted of the alleged recent offence, nor was he punished for it. He was simply assessed as to the risk which he presented.

The legal scenario which perhaps most plainly exposes the debate about the scope of article 6(2) is the civil claim for compensation made by a person who is or was a complainant in a criminal trial against the person who is or was the accused. It will of course sometimes be true that the legal constituents of the tort alleged are less exacting than those of the criminal offence (compare the disciplinary cases of Vanjak and Šikić mentioned in para 109 above). In other cases the issue in the criminal trial may be different because a defence is raised, such as mental disorder, which does not apply to a tort claim. But often the issues will be identical, and frequently the evidence relied upon will also be the same. A classic example is the claim by someone who says that she was raped by the accused. His case is either that the intercourse alleged did not take place or, more often, that it was consensual and/or that consent was to be implied from the complainant’s behaviour. He has been acquitted by the jury so it is known that the criminal standard of proof has not been achieved, but in the civil proceedings the standard is the balance of probabilities. Such cases are by no means unusual. Equally common, if not more so, are cases where a care order is sought by the Local Authority in relation to children (section 31 Children Act 1989). The test for such an order is that the child is at risk of significant harm attributable to inadequate parental care. There may be many different parental inadequacies relied upon, but a very common instance is the case which depends on an alleged risk of abuse, physical or sexual, by a parent or an associate of a parent, and where the risk is said to be proved by past abuse of this or another child. Such an alleged abuser may well also be prosecuted. If he is acquitted, on the criminal standard of proof, it is nevertheless incumbent on the family judge to investigate the allegation of past abuse in order to reach a conclusion about the level of future risk. All experienced care judges are familiar with such cases, and with the duty to find, one way or the other, on the balance of probabilities, whether the past abuse is made out despite acquittal in the criminal court.

The treatment of such cases by the Strasbourg court cannot be described as consistent. OL v Finland (Application No 61110/00), 5 July 2005 is indeed a decision that article 6(2) did not apply to child care proceedings in which one of the strands of evidence advanced concerned an allegation against the father of sexual abuse, although the prosecutor had decided not to prosecute, taking the view that the evidence was insufficient. It is perfectly true that this decision contains the proposition that article 6(2) was not applicable and that there was no link between the two sets of proceedings because the care case was “not a direct sequel” to the criminal decision. But in that case, although the psychiatrist’s report had concluded that in all likelihood the father had abused his daughter, all that the care court had said was that “it is unclear whether [the child] has been subjected to sexual abuse. This possibility cannot be excluded.” It had then gone on to record other bases for making the care order, including the disturbed behaviour of the child and the mental illness of her mother which impeded her care. If the decision as to applicability meant that the Strasbourg court took a consistent line that article 6(2) had no application to claims for civil compensation, or to care proceedings, that would be one thing. But it is clear that it does not.

In Ringvold v Norway (Application No 34964/97), 11 May 2003, the court held that article 6(2) was not applicable to the civil claim for compensation made by a victim alleging sexual abuse by an erstwhile accused who had been acquitted by the jury. It based that applicability decision in part on an absence of link (para 41) but held that this was because the outcome of the criminal proceedings was “not decisive” for the civil claim. This was to use link in an entirely different sense from the way in which it is explained in Allen at para 104. The court also based its applicability decision upon the language used in determining the civil claim (para 38). Yet it concluded that there was no applicability notwithstanding that the court had held that “on the balance of probabilities it was clear that [the erstwhile accused] was the abuser” (para 19).

Then a year later in Reeves v Norway (Application No 4248/02), 8 July 2004, the accused had been tried in the criminal courts for arson and the insurers who had paid out after the fire had been joined as civil parties to claim compensation from her. The standard of proof differed between the two decisions required, just as it would in separate proceedings in England. She was convicted at trial but on appeal her conviction for arson was quashed, on the grounds that there was not the specific majority of appeal judges which was required by local law before it could be upheld. The award of damages to the insurers was however upheld, since enough of the judges agreed that arson had been proved against her on the balance of probabilities. The Strasbourg court held that there was no infringement of article 6(2). But this time it made the assumption that article 6(2) applied to the insurers’ claim. It also found that the judgment of one judge who acquitted the appellant of the crime but found that on the balance of probabilities “there was a clear probability that the defendant is guilty of setting the fire as described in the indictment” was at risk of infringing article 6(2) and could be saved from doing so only by treating the choice of words as “an unfortunate slip” rather than as an affirmation imputing criminal liability for arson. So this decision depended not on applicability, as in Ringvold, but on whether there was infringement. The decision appears to be a good example of the unsatisfactory manner in which the language used may be determinative of whether there is a breach of article 6(2), as propounded in Allen.

Those decisions can conveniently be considered alongside Orr v Norway (Application No 31283/04), 15 May 2008, where the opposite result ensued. The accused was tried for rape of a work colleague. Her civil claim for compensation for the same rape was heard alongside the criminal trial. The jury acquitted of the crime. Next day the judges gave judgment for the complainant upon her civil claim. The applicable standards of proof differed, as they would in England, and the civil claim demanded a significantly less exacting standard, even if perhaps not a simple balance of probabilities. Giving judgment on the civil claim, the court held that on the relevant standard it was “clearly probable” that the accused had intercourse with the complainant, that it was without her consent, that he knew that it was, and that he had used sufficient force to overcome her lack of consent. The Strasbourg court did not treat a link between the criminal and civil proceedings as the test of whether article 6(2) applied or not; indeed it held that the fact that the two issues were tried together did not bring the civil part within the article. But it held that the language used did render article 6(2) applicable, and that it involved an infringement. At para 51 it held:

If, now, scenarios of this kind are tested against the Grand Chamber’s statements of principle in Allen at para 104 - for which see para 99(c) above - it would seem likely if not inevitable that article 6(2) would now be held to apply to such a civil claim for damages by a rape complainant, whether heard alongside the criminal trial or separately as it would be - and is - in any of the UK jurisdictions. The same would apply to care proceedings in which the issue was an allegation of abuse made against an acquitted accused. The judge trying such a civil claim, or such a care case, may well have to examine the evidence on the criminal file. He will certainly have to assess the accused’s participation in the events leading to the criminal charge. However, if article 6(2) really does apply to such a claim it is simply impossible for the judge in either kind of proceeding to give judgment after the accused has been acquitted. Semantic adjustment of his judgment is not an option. He has to make findings about the conflicting evidence on what occurred. He has to do so both for civil liability and to assess the level of damages. And the care judge must make findings of fact in order to justify his conclusion as to the risk of significant harm which the child faces. Neither can do other than make findings about whether the rape, or the abuse, took place. It matters not an ounce whether the judge calls it rape, or forced sexual intercourse, or abuse, and he cannot call it something which it is not. In a tort claim the tort about which he must make a finding is co-terminous in most cases with the crime; even any plea of implied consent will correlate essentially with the criminal defence of reasonable belief in consent. In a care case, it is facts constituting criminal offences which justifies the making of the care order. If article 6(2) does indeed apply to such proceedings, complainants, or public care authorities, might well consider themselves better served by not making a complaint to the police. Such allegations are notoriously difficult for juries to decide, unless there is some independent evidence beyond the word of the only two people typically present. If article 6(2) applies, an acquittal, always a possibility, will bar a finding of rape in a subsequent civil case, and thus bar the claim for compensation, and similarly with a care decision. In the absence of a prosecution, article 6(2) would presumably become irrelevant. But the public interest is unequivocally in cases of this kind being properly investigated by the police, and, if the evidence offers a reasonable prospect of conviction, in their being brought to trial.

The present case is not of course one of a civil claim for damages coming after a criminal prosecution. But consideration of such a case, together with the plain difficulties which have attended the Strasbourg court’s conscientious efforts to extend the applicability of article 6(2), demonstrates that article cannot sensibly apply beyond the criminal trial and the investigation which precedes it. The objective of not undermining an acquittal which underlies the suggested gloss on article 6(2) - see para 99(b) above - can and should properly be maintained but it means that the acquitted accused must be recognised as unconvicted, immune from punishment by the state and from characterisation as a criminal, but not that he escapes all consequences of the ordinary application of his country’s rules as to evidence and the standard of proof outside criminal trials. Powerful pleas to that effect by Judge De Gaetano in both Ashendon and Jones v United Kingdom (2012) 54 EHRR 13 and Allen, and by Judge Power in Bok v The Netherlands (Application No 45482/06), 18 January 2011, properly reflect the correct analysis of article 6(2).

This analysis of the scope of article 6(2) is, moreover, consistent with:

These same principles ought properly to govern instances where the erstwhile accused bears, under the local law, an onus of proof in proceedings which are separate from the criminal investigation and trial and in which he is at no risk of conviction or punishment. A simple example is the accused who, following acquittal which may well be in dubio pro reo brings an action for malicious prosecution against the police or other accuser. Of course it may be theoretically possible for a prosecution to be malicious even if the accused is guilty, but in most such cases it is an integral part of the claimant’s case that he was prosecuted when not guilty and that the defendant knew it. Such a claimant former accused necessarily bears the onus of proving his case, on the balance of probabilities, including his asserted innocence. No breach of article 6(2) is or could be involved, even if a link of the kind contemplated by Allen could be said to exist.

Schemes for public compensation for those who are prosecuted but acquitted vary widely from legal system to legal system. Some systems provide for compensation for detention on remand; others, including the English, have no such regime. Where there is provision for compensation, the cases show that it is not unusual for there to be some qualification to universal availability. Sometimes the system gives the court a residual discretion to withhold compensation, as for example did the Dutch scheme considered in Baars v The Netherlands (2004) 39 EHRR 25. Others state the grounds on which it may be refused, as did the German scheme considered in Nölkenbockhoff. The Strasbourg court has been at pains to say in case after case that neither article 6(2) nor any other international rule gives an unqualified right to such compensation. The limited right which is recognised internationally is that stated, in more or less identical terms, in article 3 Protocol 7 to the ECHR, for those states which have acceded to it, and in article 14(6) of the International Covenant on Civil and Political Rights. This right is limited to those whose conviction is reversed or who is pardoned, and of those only where the reversal or pardon is on the ground that a new or newly discovered fact shows conclusively that there has been a miscarriage of justice. So there is no right to compensation for those who are acquitted at trial. Nor does the right extend to the common case of a conviction quashed for error of law or of emphasis in the summing up, or for error of law, for example as to the admissibility of evidence, during the trial. Since the right to compensation is thus restricted, the test is plainly entirely different from the test of guilt or innocence at trial, and from the test of safety of the conviction on appeal. It follows, firstly, that proceedings seeking such compensation, although they are predicated upon there having been a conviction which has been quashed, so that a criminal prosecution with that outcome is a sine qua non for an award, are not part of the criminal process but rather are in aid of a distinct and limited civil right. For this reason, even if there existed a workable concept of “link” as a test for application of article 6(2), such a link would not exist between the quashing (reversal) of the conviction and the claim for compensation under section 133. The latter can only be said to be “based on” the former in the sense that the first condition of eligibility for compensation is that the conviction has been quashed. But to say that compensation is based on the quashing is to ignore the several other conditions of eligibility which must also be satisfied. Secondly it follows that it is for the claimant to show that he is within the statutory test; to that extent at least it must be common ground that he bears the onus of proof. Thirdly, it should be clear that the presumption of innocence has simply no place in such proceedings, for the simple reason that conviction and punishment are not in issue.

It is easy to understand why section 133(1ZA) can at first sight be seen as a reversal of the criminal onus of proof, and thus as inconsistent with article 6(2). In reality, however, it is no such thing. By the time section 133(1ZA) comes into consideration the erstwhile accused is by definition no longer facing any criminal charge in any sense, whether the autonomous one applied in the Strasbourg jurisprudence or any other. His conviction has been quashed. He is in no danger of conviction or punishment. Nor is he in any danger of any official body treating him as if he were still convicted or liable to punishment. All that is happening is that he is seeking to bring himself within the (legitimately) restricted eligibility requirements for compensation. That does not put his guilt or innocence in issue; he remains unconvicted and unpunished whether eligible or not, and no one will be entitled to say, if he cannot prove on the balance of probabilities that he is eligible, that he is guilty; at most all anyone could say is that his exoneration has not conclusively been proved. The terms of article 14(6) of the ICCPR, which section 133 seeks to implement in English law, make plain that eligibility depends on it being conclusively shown that a miscarriage of justice has occurred. A decision that this has not conclusively been shown is not at all the same as a finding of guilt, nor does it in any sense undermine the quashing of the conviction. As the facts of Allen show, a conviction may well be quashed on the grounds that it is not safe, without any implicit or explicit finding as to guilt or innocence: see Allen at paras 127, 131-132 and 134-135. An English lawyer might baulk at the assertion in para 127 that the appellant in that case had not been “acquitted on the merits” since he or she would say that a decision that the conviction is unsafe is indeed a judgment on the merits, but the sense of the court’s judgment is clear: those adjudicating on the question of compensation “did not comment on whether, on the basis of the evidence as it stood at the appeal, the applicant should be, or would likely be, acquitted or convicted. Equally, they did not comment on whether the evidence was indicative of the applicant’s guilt or innocence.” (para 134). This will be equally true whenever a claimant, seeking compensation after the enactment by Parliament of section 133(1ZA) fails the eligibility test which it creates.

This court’s obligation under section 2(1)(a) of the Human Rights Act 1998 is to “take into account” any judgment, decision, declaration or advisory opinion of the Strasbourg court. Its ultimate responsibility is to arrive at its own decision on those Convention rights which are given domestic legal effect by being incorporated into that statute. The history of the English courts rightly demonstrates a desire if at all possible to maintain consistency of approach with the Strasbourg court. That desire is reflected in the general proposition that an English court “should in the absence of some special circumstances, follow any clear and constant jurisprudence of the Strasbourg court”: R (Ullah) v Special Adjudicator [2004] UKHL 26; [2004] 2 AC 323, para 20. I respectfully share that desire and the present judgment sets out to take account of the Strasbourg jurisprudence in some detail.

In accordance with its usual practice, the Strasbourg court has often repeated, usually in identical language, the key propositions which are once again propounded in Allen and which are summarised at para 99 above. To the extent that they are oft repeated, they are no doubt “constant”. To say that they are clear is, on inspection, much more difficult. As appears from the brief survey above they create considerable difficulties in application, frequently leading either to inconsistent outcomes or to over-sophisticated semantic analysis in an effort to achieve the right result. It seems to me incumbent on this court to stand back and re-assess the a priori assumption that has been made that the presumption of innocence, and the critical requirement to respect acquittals or reversals of convictions, extends to preventing any comment by any court “in any other proceedings of whatever nature” (Vanjak)which assesses conduct which was in question also in the criminal proceedings. Proper respect for acquittal does not require this. It requires that the erstwhile accused is treated as acquitted, not that his conduct cannot fall for examination in other proceedings where the test is quite different from the criminal standard of proof.

For these reasons, which substantially although not exactly overlap with those of Lords Mance and Wilson, I would dismiss these appeals. The correct analysis is that article 6(2) does not apply to section 133 claims for compensation. It certainly requires that in such claims, as in any other proceedings, the reversal of the conviction is treated as unquestioned. But it does not inject into the quite different section 133 test a presumption that the erstwhile accused did not commit the crime; it holds that he has not been proved to the strict criminal standard to be guilty. Nor therefore does article 6(2) apply so as to strike down the provision in section 133(1ZA) which makes clear that a claimant for compensation must accept the onus of bringing himself within the eligibility criteria laid down by Parliament.

If, contrary to that clear view, it be held that this court is duty bound by the Strasbourg jurisprudence to hold that article 6(2) does apply to a section 133 claim, I would conclude with the Court of Appeal and Divisional Court below that to require a claimant to prove his case of eligibility is not a breach of it. That is because what article 6(2) (if it applies) preserves is the presumption of innocence in the sense of being a person who is acquitted, unconvicted and unpunishable. “Innocence”, in the context of the criminal law and of article 6(2), does not mean “exonerated on the facts”; it means “unconvicted, not proved according to the governing standard of proof, accordingly not liable to punishment, and entitled to be treated as such”. The new section 133(1ZA) does not require the claimant to prove that he has this status. This status (which appears to be what the courts below meant by “innocence” in a general sense) is already a given, once the conviction has been quashed by the Court of Appeal (Criminal Division). What the new section requires is that the claimant prove something different and additional, viz the condition of eligibility for compensation under the scheme established in England and Wales. I agree that the mere fact that the section requires exoneration as a result of a new or newly discovered fact would not prevent it from calling for proof of innocence, or from conflicting with the presumption of innocence, if “innocence” in the context of the presumption meant “exonerated on the facts”. But for the reasons explained, it does not and cannot. The difference is clearly stated by Sir Thomas Bingham MR in R v Secretary of State for the Home Department, Ex p Bateman (1994) 7 Admin LR 175, cited by Lord Dyson MR in the Court of Appeal below at para 49.

This critical distinction between “innocence” as used in article 6(2) and exoneration on the facts might in one sense be said to be a semantic one, but if so the Strasbourg court has emphasised time and again that language (ie semantics) is for it the critical test of breach of article 6(2). In reality it is not a mere semantic distinction but reflects a fundamental principle of the criminal law, namely the strict enhanced standard of proof. It is not possible for the law simultaneously to erect a differential and enhanced standard of proof for criminal prosecutions, and then effectively to apply that standard not just to criminal trials but to other (indeed maybe to all) other adjudications upon the facts which led to the prosecution. Neither the suggested test of “link” nor the suggested test of language will work to determine the scope of article 6(2) in the face of this difficulty.

The form of judgments in the Court of Appeal when dealing with appeals against conviction is not the issue in the present case. It is, however, important that that court is not constrained in giving its reasons either for dismissing an appeal or for allowing it. I do not disagree with what Lord Mance says at paras 25-34. In summary:

I agree with the judgment of Lord Mance and therefore limit myself to some brief observations on the position which has been reached in the Strasbourg jurisprudence in relation to the scope of application of article 6(2) ECHR after acquittal or discontinuance of criminal proceedings.

I agree with Lord Mance’s analysis of the case law of the ECtHR. For the reasons he gives, I too would decline to follow that case law if and to the extent that it may have gone beyond precluding reasoning that suggests that a defendant in criminal proceedings leading to an acquittal or discontinuance should have been convicted of the criminal offence with which he was charged.

In any event, I consider that the incompatibility of section 133(1ZA) with article 6(2) is not made out. The objection to the section as amended is, as I understand it, that it requires the Secretary of State to assess whether persons whose convictions are quashed because of fresh evidence have established by that evidence that they are innocent.

The Strasbourg case law makes clear that there is nothing objectionable in resisting or refusing compensation on the ground that the case falls within category (3) ie where fresh evidence renders the conviction unsafe in that, had it been available at the time of the trial, a reasonable jury might or might not have convicted the defendant. See ALF v United Kingdom (Application No 5908/12), 12 November 2013. That is also apparent from Allen, a category (3) case where it was not suggested that the case fell into a higher category.

It must also follow from Allen that there is nothing objectionable in requiring a claimant seeking compensation to bring himself or herself within category (2) ie where the fresh evidence so undermines the evidence against the defendant that no conviction could possibly have been based upon it. This is the whole thrust of the decision in Allen. By the same token, there can be nothing objectionable in the state contending against such an outcome in the circumstances of a particular case.

Yet it seems that the line is drawn in the Strasbourg case law at requiring a claimant to demonstrate his or her innocence ie to bring himself or herself within category (1), where the fresh evidence shows clearly that he or she is innocent of the crime. This is apparent from the observation in Allen (at para 133) that “what is important above all is that the judgments of the High Court and the Court of Appeal did not require the applicant to satisfy Lord Steyn’s test [in Mullen] of demonstrating her innocence”.

The difficulty with this approach, as Lord Mance points out, is that category (2) subsumes category (1). It is, no doubt, possible to draw a distinction between category (1) and category (2) but I am, at present, unable to see why this should be significant in the present context. I can see no sensible basis on which it is held objectionable to require evidence which establishes innocence but not objectionable to require evidence which establishes that the claimant could not reasonably have been convicted. Moreover - and to this I attach particular importance - this specific issue has not yet been directly addressed or decided by the ECtHR.

Having regard to the present unsettled state of ECtHR case law, therefore, I am not persuaded that section 133(1ZA) is incompatible with article 6(2). It seems to me that these are matters which require consideration by the ECtHR and which that court will be anxious to address.

For these reasons I would refuse declarations of incompatibility and would dismiss the appeals.

I am grateful to Lord Mance for setting out the background to these appeals and the issues arising.

The terms of article 6(2) of the European Convention on Human Rights are set out in para 35 above. Read literally, the words “charged with a criminal offence” might suggest that the guarantee only applies in the context of pending criminal proceedings. But it has never been interpreted so narrowly. In the first place, the European court long ago adopted the position that the character of a procedure under domestic law cannot be decisive of the question whether article 6 is applicable, since the guarantees contained in that provision could otherwise be avoided by the classification of proceedings. The case law on article 6(1) has therefore made it clear that the concept of a “criminal charge” has an autonomous meaning, with the consequence that article 6(2) is applicable to proceedings which may not be classified as criminal under domestic law, provided that they satisfy the criteria developed in cases such as Engel v The Netherlands (No 1)(1976) 1 EHRR 647 and Öztürk v Germany(1984) 6 EHRR 409. Secondly, it has also long been clear from the case law of the European court that the scope of article 6(2) is not limited to pending criminal proceedings as so defined, but extends in some circumstances to decisions taken by the state after a prosecution has been discontinued or after an acquittal.

The case law of the European court concerning the scope of article 6(2), prior to the judgment of the Grand Chamber in Allen v United Kingdom (2013) 63 EHRR 10, was considered by this court in the case of R (Adams) v Secretary of State for Justice (JUSTICE intervening) [2011] UKSC 18; [2012] 1 AC 48. The implication of the court’s decision in that case is that article 6(2) has no application to section 133 of the Criminal Justice Act 1988 (“the 1988 Act”). The first question which arises in this appeal is whether this court should follow that decision, as the Secretary of State submitted, or should depart from it, as the appellants invited us to do, in the light of the decision in Allen v United Kingdom that article 6(2) applies to decisions taken under section 133.

The judgments in Adams did not differentiate clearly between the question whether article 6(2) is applicable and the question whether it has been infringed. As a consequence, it is difficult to be certain which of the arguments accepted by the court were thought to bear on the former question, and which were concerned with the latter. The fullest analysis was carried out by Lord Hope, who based his conclusion at para 111 that article 6(2) had no “impact” on section 133 on three arguments, which were also advanced on behalf of the Secretary of State in the present proceedings. They can be discussed under the headings (a) lex specialis, (b) separate proceedings, and (c) not undermining the acquittal. It is necessary to consider each of these in turn.

Lord Hope considered that article 6(2) and article 3 of Protocol No 7 (“A3P7”) stood in the relation of lex generalis and lex specialis respectively, so that the maxim lex specialis derogat legi generali applied: that is to say, that where a legal issue falls within the ambit of a provision framed in general terms, but is also specifically addressed by another provision, the specific provision overrides the more general one. This was, with respect, a questionable conclusion, since article 6(2) and A3P7 are concerned with different issues: article 6(2) is concerned with the presumption of innocence, whereas A3P7 is concerned with the payment of compensation to persons whose convictions have been quashed, and is silent about the presumption of innocence. Since they concern different issues, they are capable of applying cumulatively, rather than it being necessary to apply one to the exclusion of the other.

Lord Hope found support for the view that the maxim applied in the speech of Lord Steyn in R (Mullen) v Secretary of State for the Home Department [2004] UKHL 18; [2005] 1 AC 1. Referring to article 14(6) of the ICCPR, set out in para 16 above, and to article 14(2) (“Everyone charged with a criminal offence shall have the right to be presumed innocent until proved guilty according to law”), Lord Steyn cited the report of the UN Human Rights Committee in WJH v The Netherlands (Communication No 408/1990) [1992] UNHRC 25, where the Committee said at para 6.2:

Whatever the merits of that view may be in relation to the ICCPR, it might be doubted whether it is of assistance in deciding the scope of article 6(2) of the Convention, since it depends on the Human Rights Committee’s statement that article 14(2) of the ICCPR applies only to criminal proceedings and not to proceedings for compensation. Whether that is true of article 6(2) is the very question in issue. In relation to that question, although Lord Steyn cited a number of European cases, such as Sekanina v Austria(1993) 17 EHRR 221, which demonstrated that article 6(2) could apply to proceedings for compensation, he concluded at para 44 that “the European jurisprudence cited throws no light on the question”, and that “article 14(6) of the ICCPR (and therefore section 133 of the 1988 Act), are in the category of lex specialis and the general provision for a presumption of innocence does not have any impact on it”. This analysis might be contrasted with that of Lord Bingham, who pointed out at para 10 that the fact that article 6(2) was not confined to criminal proceedings, as illustrated by Sekanina, indicated that the European court took a different approach from that taken by the Human Rights Committee in relation to article 14(2) of the ICCPR.

In support of his conclusion, Lord Steyn also referred to the Explanatory Report to Protocol No 7, prepared by the Steering Committee for Human Rights appointed by the Council of Europe. In relation to A3P7, the report stated at para 25:

The question whether section 133 of the 1988 Act fell within the ambit of article 6(2) of the Convention did not, however, have to be decided in Mullen. Lord Hope returned to it in Adams. He accepted Lord Bingham’s reasons for doubting whether Lord Steyn was right to find support for his view in the French text and in para 25 of the Explanatory Report, and therefore took a fresh look at the issue. His conclusion that section 133 fell outside the ambit of article 6(2) was based, as explained above, on the view that article 6(2) was excluded from applying within the scope of A3P7, since the latter was lex specialis relative to the lex generalis contained in the former. In forming that view, he relied on a passage in the court’s judgment in Sekanina, in the section dealing not with applicability but with compliance. After explaining that “article 6(2) does not guarantee a person ‘charged with a criminal offence’ a right to compensation for detention on remand”, the European court added at para 25:

As explained above, A3P7 requires the payment of compensation to a person who has suffered punishment as a result of a conviction which is subsequently reversed on the ground that a new or newly discovered fact shows conclusively that there has been a miscarriage of justice. As the court stated in para 25 of Sekanina, the situation of the applicant in that case was not comparable to that governed by A3P7: he was seeking compensation for having been remanded in custody pending a trial at which he was acquitted, whereas A3P7 applies to persons who have suffered punishment as a result of a conviction. That is all that the court said in the relevant passage. Lord Hope, however, read more into it, stating at para 111:

The second strand in Lord Hope’s reasoning concerned the relationship between the determination of a claim under section 133 of the 1988 Act and the antecedent criminal proceedings. He stated at para 109 that “the Strasbourg cases show that its jurisprudence is designed to protect the criminal acquittal in proceedings that are closely linked to the criminal process itself”, and went on at para 111 to distinguish “comments on the underlying facts of the case in subsequent proceedings of a different kind, such as a civil claim for damages”. He illustrated the point by reference to Sekanina, noting that in its judgment the court said at para 22 that the Austrian legislation and practice linked “the criminal responsibility of the accused and the right to compensation - to such a degree that the decision on the latter issue can be regarded as a consequence and, to some extent, the concomitant, of the decision on the former”. Lord Hope concluded that the system laid down by article 14(6) of the ICCPR, and implemented by section 133, did not cross the forbidden boundary, stating at para 111:

This reasoning is also questionable. Although procedurally separate, compensation proceedings under section 133 are nevertheless based on the quashing of a conviction by the criminal courts, and are directed towards obtaining compensation for harm inflicted by the state as a direct consequence of that conviction. But for the outcome of the criminal proceedings, there could be no compensation proceedings. In the language used by the European court, the outcome of the criminal proceedings is therefore “decisive” for the compensation proceedings, since it is a prerequisite of a compensation claim that the conviction has been quashed. The time limit for bringing a claim is also directly linked to the conclusion of the criminal proceedings: a factor which was regarded as relevant in a series of cases concerned with compensation proceedings under Norwegian law, such as Hammern v Norway (Application No 30287/96) (unreported) given 11 February 2003, para 43. Furthermore, the decision whether to award compensation, even before the amendment of section 133, depended on an assessment of the circumstances in which the conviction was quashed, based on an examination and evaluation of the judgment of the Court of Appeal. In these circumstances, even prior to Allen v United Kingdom, the Strasbourg case law clearly indicated that the compensation proceedings were likely to be regarded as a sequel or, as it was put in Sekanina, a consequence and concomitant, of the criminal proceedings, and therefore within the ambit of article 6(2).

Finally, Lord Hope considered that a refusal of compensation under section 133, prior to its amendment, did not have the effect of undermining the acquittal in the criminal proceedings. That conclusion is consistent with that of the European court in Allen v United Kingdom and later cases. However, it goes to the question whether article 6(2) has been violated, not to the question whether it is applicable.

Lord Phillips and Lord Kerr agreed with Lord Hope on this topic. Lord Judge CJ, with whom Lord Brown, Lord Rodger and Lord Walker agreed on this topic, also treated A3P7 as a lex specialis which ousted the application of article 6(2) to proceedings under section 133. In the present case, the courts below were therefore correct to take the view that they were bound by Adams to hold that article 6(2) was inapplicable.

Before turning to the more recent Strasbourg jurisprudence, it is also relevant to note the case of Serious Organised Crime Agency v Gale (Secretary of State for the Home Department intervening)[2011] UKSC 49; [2011] 1 WLR 2760, decided by this court a few months after Adams. The case concerned the question whether civil recovery proceedings under the Proceeds of Crime Act 2002, undertaken following the appellant’s acquittal of criminal charges, were compatible with article 6(2). In the course of his judgment, with which a majority of the court agreed, Lord Phillips was critical of the distinction which he perceived in the case law of the European court between claims for compensation brought by an acquitted defendant against the state under public law, and claims for compensation brought by an alleged victim against an acquitted defendant under the law of tort, commenting at para 32 that “this confusing area of Strasbourg law would benefit from consideration by the Grand Chamber”. Lord Dyson was less critical of the Strasbourg jurisprudence, and provided an illuminating analysis.

As he noted, cases in which article 6(2) was held to apply to proceedings instituted after the discontinuation of criminal proceedings or following an acquittal included, first, cases in which there was a sufficiently close link between the criminal proceedings and the other proceedings to engage article 6(2), even if on an application of the usual Engel criteria the latter proceedings would be characterised as civil. Those cases were described in Ringvold v Norway Reports of Judgments and Decisions 2003-II, p 117, para 36 as concerning “proceedings relating to such matters as an accused’s obligation to bear court costs and prosecution expenses, a claim for reimbursement of his (or his heirs’) necessary costs, or compensation for detention on remand, matters which were found to constitute a consequence and the concomitant of the criminal proceedings”. The focus of the inquiry was on whether the proceedings were the “direct sequel” or “a consequence and the concomitant” of the criminal proceedings: ibid, at para 41. As Lord Dyson stated at para 125:

As Lord Dyson explained, civil claims for compensation, brought against the defendant under the law of tort, are not linked in that way to criminal proceedings. The victim of a civil wrong has a right to claim damages, in order to obtain a remedy for the harm which he or she has suffered, regardless of whether the defendant has been convicted or acquitted of a criminal offence arising out of the same facts. The victim’s claim is not dependent on the defendant being prosecuted at all. Furthermore, as the court pointed out in Ringvold, para 38, if civil compensation proceedings automatically fell within the ambit of article 6(2), that would have:

A separate basis on which article 6(2) had been held to apply to proceedings instituted after the discontinuation of criminal proceedings or following an acquittal was that a sufficient link with the criminal proceedings was created by the language used by the court in the civil proceedings. An example was the case of Y v Norway(2003) 41 EHRR 87, where the civil court stated in its judgment that it found it clearly probable that the defendant had committed the offences against the claimant with which he was charged. The European court found that there had been a violation of article 6(2).

Lord Dyson contrasted that case with Moullet v France (Application No 27521/04) (unreported) given 13 September 2007, where the applicant was a public official who had been charged with accepting bribes. The criminal proceedings were discontinued on the ground that they were time-barred. The official was then dismissed on the basis that the evidence showed that he had taken bribes. That decision was challenged under administrative law, but was upheld by the Conseil d’Etat on the ground that it had been based on “accurate facts” and on reasons which were not “materially or factually incorrect”. A complaint to the European court was unsuccessful. The court considered whether the Conseil d’Etat “used such language in its reasoning as to create a clear link between the criminal case and the ensuing administrative proceedings and thus to justify extending the scope of article 6(2) to cover the latter”. It noted that the applicant was not “formally declared guilty of the criminal offence of accepting bribes”. The Conseil d’Etat had confined itself to determining the facts “without suggesting any criminal characterisation whatsoever … In other words, the domestic authorities managed in the instant case to keep their decision within a purely administrative sphere, where the presumption of innocence the applicant relied on did not obtain”.

Similarly in Ringvold v Norway the court found that a domestic decision awarding compensation to a victim of sexual abuse, following the defendant’s acquittal, did not fall within the scope of article 6(2). Although the domestic court had found that there was evidence “establishing that sexual abuse had occurred, and that, on the balance of probabilities, it was clear that the applicant was the abuser” (para 19), it “did not state, either expressly or in substance, that all the conditions were fulfilled for holding the applicant criminally liable with respect to the charges of which he had been acquitted” (para 38).

Lord Dyson commented at para 138 that the rationale of cases such as Y v Norway must be that if the domestic court chooses to treat civil proceedings as if the issue of criminal liability falls to be determined, then the fair trial protections afforded by article 6(2) should be respected. But if the decision in the civil proceedings is based on reasoning and language which go no further than is necessary for the purpose of determining the issue before that court and without making imputations of criminal liability, then the necessary link will not have been created.

An opportunity for the Grand Chamber to consider this area of the law arose soon after Gale, in the case of Allen v United Kingdom. The applicant had been convicted of manslaughter. Her conviction was later quashed on the basis that, although the Crown case against her remained strong, a jury which had heard the fresh evidence might have come to a different conclusion. In terms of the categories subsequently adopted in Adams, it was a category 3 case. Her application for compensation under section 133 as originally enacted was unsuccessful, and her application for judicial review of that decision was dismissed. On appeal, the Court of Appeal held that there had been no violation of article 6(2): R (Allen) (formerly Harris) v Secretary of State for Justice[2008] EWCA Civ 808; [2009] 1 Cr App R 2. As was pointed out, article 6(2) could not possibly mean that compensation necessarily followed the quashing of a conviction on the basis of fresh evidence, otherwise A3P7 could not be in the terms it was. More controversially, Hughes LJ, giving the judgment of the court, expressed the view, applying dicta of Lord Steyn in the case of Mullen, that the phrase “miscarriage of justice” in section 133 of the 1988 Act was restricted to cases where the defendant was demonstrably innocent of the crimes of which he had been convicted: a view which was subsequently disapproved by the majority of this court in Adams.

When Allen reached the Grand Chamber of the European court, on a complaint directed not against the Secretary of State’s decision to refuse the applicant’s claim for compensation, but against the reasons given by the High Court and the Court of Appeal for dismissing her challenge to that decision, the European court was therefore considering section 133 in its unamended form. The Government contended that the complaint was inadmissible because article 6(2) had no application to decisions taken under section 133, as this court had held in Adams. The question whether section 133 fell within the scope of article 6(2) was therefore directly in issue. In deciding that question, the Grand Chamber court undertook a careful review of the court’s case law, and considered the relationship between article 6(2) and A3P7.

The Grand Chamber began its assessment by explaining the justification, in accordance with the most fundamental principles of the Convention case law, for giving article 6(2) a wider application than a literal reading of the text would suggest. As it explained at para 92:

The Grand Chamber reviewed how the court’s jurisprudence in relation to the second aspect of article 6(2) had developed over time. In doing so, it did not attempt to justify or reconcile all of the decisions on their particular facts: a task which, in relation to some of the case law, might have been challenging. Instead, it sought to derive from the cases the underlying principles, and to explain how they had evolved. In some early cases in which the court had found article 6(2) to be applicable, despite the absence of a pending criminal charge, it had said that the judicial decisions taken following criminal proceedings, for example with regard to an obligation to bear court and prosecution costs, or compensation for pre-trial detention or other adverse consequences, were “consequences and necessary concomitants of”, or “a direct sequel to”, the conclusion of the criminal proceedings. Similarly, in a later series of cases, such as Sekanina v Austria, it had concluded that Austrian legislation and practice “link[ed] the two questions - the criminal responsibility of the accused and the right to compensation - to such a degree that the decision on the latter issue could be regarded as a consequence and, to some extent, the concomitant of the decision on the former”, so that article 6(2) applied to the compensation proceedings. Developing this idea in subsequent cases, such as Hammern v Norway, the court had found that the applicants’ compensation claim “not only followed the criminal proceedings in time, but was also tied to those proceedings in legislation and practice, with regard to both jurisdiction and subject matter”, creating a link between the two sets of proceedings with the result that article 6(2) was applicable. In cases such as Ringvold v Norway and Y v Norway, concerning the victim’s right to compensation from the applicant, who had previously been found not guilty of the criminal charge, the court had held that where the decision on civil compensation contained a statement imputing criminal liability, this would create a link between the two proceedings such as to engage article 6(2) in respect of the judgment on the compensation claim.

The Grand Chamber also cited its decision in OL v Finland (Application No 61110/00) (unreported) given 5 July 2005, in which an appeal was brought against a child care order, made on the basis of a psychiatric report stating that it was highly probable that the child had been sexually abused by her father, after the public prosecutor decided not to bring charges. In dismissing the appeal, the domestic court stated:

More recently, the court had expressed the view that following the discontinuation of criminal proceedings, the presumption of innocence required that the lack of a person’s criminal conviction should be preserved in any other proceedings of whatever nature. It had also indicated that the operative part of an acquittal judgment must be respected by any authority referring directly or indirectly to the criminal responsibility of the person in question.

The Grand Chamber then considered the specific context of judicial proceedings following the quashing of a conviction, giving rise to an acquittal, and stated at para 104:

The Grand Chamber next addressed the argument that article 6(2) did not apply to section 133 of the 1988 Act because the latter fell within the scope of A3P7, which was argued to be lex specialis: the argument accepted by a majority of this court in Adams. The Grand Chamber had earlier mentioned the UN Human Rights Committee’s communication in WJH v The Netherlands, which, as it noted, proceeded on the basis that article 14(2) of the ICCPR applied only to criminal proceedings. It also cited the Explanatory Report on Protocol No 7, including the passages to which Lord Bingham had referred in Mullen, observing at para 133 that the report itself provided that it did not constitute an authoritative interpretation of the text, and adding that the report’s reference to the need to demonstrate innocence must now be considered to have been overtaken by the court’s intervening case law on article 6(2). It concluded at para 105:

The Grand Chamber then applied the general principles set out earlier in its judgment to the facts of Allen. It identified the relevant question as being “whether there was a link between the concluded criminal proceedings and the compensation proceedings, having regard to the relevant considerations” set out in para 104 of the judgment. In that regard, it stated at paras 107-108:

The critical factors in establishing the necessary link between the decision of the Court of Appeal in the criminal proceedings, and the subsequent proceedings under section 133, were therefore that the quashing of the conviction was a prerequisite of proceedings under section 133, and that in order to arrive at a decision on the claim it was necessary for the Secretary of State to examine the judgment of the Court of Appeal so as to determine whether the criteria in section 133 were satisfied. That reasoning applies equally, if not a fortiori, to section 133 in its amended form.

The only remaining question, therefore, in relation to the applicability of article 6(2) to decisions taken under section 133 as amended, is whether, as counsel for the Secretary of State submitted, this court should decline to follow the decision of the Grand Chamber. In counsel’s submission, our doing so would encourage, or stimulate, further dialogue where the issue could be reviewed and addressed in full.

This court’s approach to judgments of the European Court of Human Rights is well established. Section 2 of the Human Rights Act requires the courts to “take into account” decisions of the European court, not necessarily to follow them. In taking them into account, this court recognises their particular significance. As Lord Bingham observed in Kay v Lambeth London Borough Council[2006] UKHL 10; [2006] 2 AC 465, para 44:

The circumstances in which constructive dialogue is realistically in prospect are not, however, unlimited. As Lord Neuberger of Abbotsbury MR explained in Manchester City Council v Pinnock (Secretary of State for Communities and Local Government intervening)[2010] UKSC 45; [2011] 2 AC 104, para 48:

No circumstances of the kind contemplated in those dicta exist in the present case. The Grand Chamber’s conclusion was carefully considered, and was based on a detailed analysis of the relevant Strasbourg case law. It was consistent with a line of authorities going back decades. It was intended to provide authoritative guidance, and has been followed in numerous subsequent judgments, such as Cleve v Germany (Application No 48144/09) (unreported) given 15 January 2015, Kapetanios v Greece (Application Nos 3453/12, 42941/12 and 9028/13) (unreported) given 30 April 2015 and Dicle and Sadak v Turkey (Application No 48621/07) (unreported) given 16 June 2015. It did not involve any principle of English law, or any oversight or misunderstanding. On the contrary, it is the reasons given in Adams to support the conclusion that article 6(2) has no application to section 133 of the 1988 Act which, with respect, are less than compelling. The lex specialis argument is unpersuasive, for the reasons explained at paras 144-149 above, and those set out by the Grand Chamber at para 105 of its judgment. The “separate proceedings” argument is equally unpersuasive, as explained at para 151 above, and at para 107 of the Grand Chamber’s judgment. That is also the implication of Lord Dyson’s analysis in Gale, where he explained at para 125 (quoted in para 155 above) why claims by a defendant for compensation for detention are a paradigm example of proceedings which are sufficiently closely linked to criminal proceedings for article 6(2) to apply. The “not undermining the acquittal” argument bears on compliance with article 6(2), not on whether it is applicable.

I recognise that the dicta which I have cited from Pinnock and Chester are not to be treated as if they had statutory force. Nevertheless, they are in my view persuasive. I find it difficult to accept that this court should deliberately adopt a construction of the Convention which it knows to be out of step with the approach of the European Court of Human Rights, established by numerous Chamber judgments over the course of decades, and confirmed at the level of the Grand Chamber, in the absence of some compelling justification for taking such an exceptional step. For my part, I can see no such justification.

For the reasons I have explained, I would hold that decisions taken under section 133 fall within the ambit of article 6(2). I would therefore depart from the decision in Adams in so far as it adopted the contrary view.

Once it has been established that there is a sufficient link between proceedings under section 133 and the antecedent criminal proceedings, the court must determine whether the presumption of innocence has been respected. The approach to be adopted to this question was the second area of the law which was reviewed by the Grand Chamber in Allen v United Kingdom.

As the court observed, there is no single approach to ascertaining the circumstances in which article 6(2) will be violated in the context of proceedings which follow the conclusion of criminal proceedings. In particular, the court explained in para 121 that in cases concerning applications by a former accused for compensation or costs, where the criminal proceedings were discontinued, it had been held that a refusal of compensation or costs might raise an issue under article 6(2) “if supporting reasoning which could not be dissociated from the operative provisions amounted in substance to a determination of the accused’s guilt”, but that no violation had been found where domestic courts had described a “state of suspicion” without making any finding of guilt. In Sekanina, however, the court drew a distinction between cases where the criminal proceedings had been discontinued and those where a final acquittal judgment had been handed down

To give one other example, in cases involving civil compensation claims lodged by victims, regardless of whether the criminal proceedings ended in discontinuation or acquittal, the court had emphasised that while exoneration from criminal liability ought to be respected in the civil compensation proceedings, it should not preclude the establishment of civil liability to pay compensation arising out of the same facts on the basis of a less strict burden of proof. However, if the national decision on compensation were to contain a statement imputing criminal liability to the respondent party, this would raise an issue falling within the ambit of article 6(2).

Turning to consider the circumstances in Allen itself, the court observed that the applicant’s conviction was quashed on the ground that it was “unsafe”, because new evidence might have affected the jury’s decision had it been available at trial. The Court of Appeal did not itself assess all the evidence in order to decide whether guilt had been established beyond reasonable doubt. Nor had it ordered a retrial, since the applicant had already served her sentence. In these circumstances, although the quashing of the conviction resulted in a verdict of acquittal being entered, it was not “an acquittal ‘on the merits’ in a true sense”. In that respect, the court contrasted the case with Sekanina and the similar case of Rushiti v Austria(2001) 33 EHRR 56, “where the acquittal was based on the principle that any reasonable doubt should be considered in favour of the accused”. The court observed, at para 127, that

The court next considered whether the criteria laid down by section 133 as originally enacted were themselves incompatible with article 6(2). As it observed, there was nothing in the criteria which called into question the innocence of an acquitted person, and the legislation did not require any assessment of the applicant’s criminal guilt.

The court next considered the approach adopted by the domestic courts in the case before it. They had been entitled under the Convention to conclude that more than an acquittal was required in order to establish a miscarriage of justice, “provided always that they did not call into question the applicant’s innocence”. In that regard, the court referred to the view expressed by Lord Steyn in Mullen (subsequently adopted by the minority in Adams) that a miscarriage of justice, within the meaning of section 133(1), would only arise where the person concerned was innocent, and that section 133 therefore required that the new or newly discovered fact must demonstrate the applicant’s innocence beyond reasonable doubt. The court observed that “what is important above all is that the judgments of the High Court and the Court of Appeal did not require the applicant to satisfy Lord Steyn’s test of demonstrating her innocence”.

The difference in the present case is that the insertion of section 133(1ZA) into the 1988 Act has had the effect of introducing a test that the fresh evidence has to establish beyond reasonable doubt that the applicant did not commit the offence. In the present proceedings, the Divisional Court and the Court of Appeal considered this test to be compatible with article 6(2), since it did not require the applicant to establish his innocence, but imposed a narrower requirement, namely that he demonstrate that his innocence had been established by a new or newly discovered fact “and nothing else”, as the Court of Appeal stated at para 48. The refusal of an application under section 133 did not, therefore, in their view cast doubt on the person’s innocence generally. The Court of Appeal observed that a focus on the new or newly discovered fact and nothing else was central to limiting eligibility for compensation to a narrower category of cases than the entire corpus of cases where a conviction was quashed. It also considered that the European court’s observations about Lord Steyn’s test in Mullen were directed to the dangers of imposing a general requirement of having to demonstrate innocence, which was not what was required by section 133.

I do not find this an easy question, but I have respectfully come to a different conclusion from the courts below. In the context of decisions made under the amended section 133, the distinction between a requirement that innocence be established, and a requirement that innocence be established by a new or newly discovered fact and nothing else, appears to me to be unrealistic. A person who can make a valid application under section 133 is, of necessity, someone whose conviction has been quashed because of the impact of a new or newly discovered fact: that follows from the terms of section 133(1). In most cases which satisfy that criterion, there will not be any other reason for the quashing of the conviction. A decision by the Secretary of State that the new or newly discovered fact does not establish the person’s innocence does not, therefore, usually leave open a realistic possibility that he or she has been acquitted for some other reason, which that decision leaves unaffected. On the contrary, the implication of the decision is likely to be that, although the new or newly discovered fact has led to the quashing of the conviction, the person’s innocence has not been established. The decision therefore casts doubt on the innocence of the person in question and undermines the acquittal.

The idea that there is a meaningful distinction between assessing whether innocence has been established by a new or newly discovered fact, and assessing whether innocence has been established in a more general sense, also appears to me to be unrealistic for another reason. Normally, at least, the significance of a new piece of evidence can only be assessed in the context of the evidence as a whole. That is illustrated by the present cases. The photograph of Mr Hallam in Mr Harrington’s company does not in itself tell one anything about his guilt or innocence of the murder. It is only when considered in the context of the alibi evidence that its significance becomes apparent. In Mr Nealon’s case, the presence of an unknown male’s DNA on the victim’s underwear tells one nothing in itself about Mr Nealon’s guilt or innocence of an attempted rape. It is only in the context of her evidence about the behaviour of her attacker and her contact with other males on the day in question, and the evidence of other witnesses eliminating the most likely alternative explanations of the presence of the DNA, that its significance can be assessed. There is no material difference, in these situations, between asking whether the applicant’s innocence has been established by the new or newly discovered fact, and asking whether his innocence has been established.

The majority of this court have reached the same conclusion as the courts below, but for somewhat different reasons. As I understand their reasoning, they emphasise that, in Allen v United Kingdom, the Grand Chamber found no violation of article 6(2) in the judgment of the Court of Appeal upholding the refusal of compensation under section 133 in its original form to an applicant who, in terms of the domestic categories subsequently adopted in Adams, fell into category 3, and failed to fall into category 2. They consider that it must, or at least may, be equally compatible with article 6(2) to require the applicant to demonstrate that he falls into category 1.

I accept that the implication of the decision in Allen v United Kingdom is that it is not necessarily incompatible with article 6(2) to refuse compensation under section 133 in cases falling within the category later described in Adams as category 3: that is to say, cases where the fresh evidence renders the conviction unsafe in that, had it been available at the time of the trial, a reasonable jury might or might not have convicted the defendant. The effect of the decision of this court in Adams, confining compensation to cases in category 2 (where the fresh evidence so undermines the evidence against the defendant that no conviction could possibly be based upon it), has been held in later cases before the European court to be compatible with article 6(2): see, for example, ALF v United Kingdom (Application No 5908/12) (unreported) given 12 November 2013. It is not a violation of the presumption of innocence to say that a case falling within category 3 (or category 4: cases where something has gone seriously wrong in the investigation of the offence or the conduct of the trial, resulting in the conviction of someone who should not have been convicted) does not constitute a miscarriage of justice. Nor is there any objection under article 6(2) to other criteria for the award of compensation that do not require the applicant to establish his or her innocence: for example, criteria precluding compensation where successful appeals are brought within time, or where convictions are quashed because of misdirections. The problem which arises under article 6(2) when compensation is confined to persons in category 1 - cases where the fresh evidence shows clearly that the defendant is innocent of the crime of which he was convicted - as under section 133 as amended, is quite specific. It is that it effectively requires the Secretary of State to decide whether persons whose convictions are quashed because of fresh evidence have established that they are innocent. In Allen, the Grand Chamber found at para 128 that there was nothing in the criteria set out in section 133 as it then stood which called into question the innocence of an acquitted person, and that the legislation itself did not require any assessment of the applicant’s criminal guilt. I doubt whether the same could be said of section 133 in its amended form.

In cases falling within category 2, the person has received an acquittal “on the merits”, in the language used by the European court: the Court of Appeal has assessed all the evidence and has concluded that, allowing the defendant the benefit of any reasonable doubt, only a verdict of acquittal could reasonably be arrived at. The principle in Sekanina therefore applies, and it is no longer permissible to rely on suspicions regarding the defendant’s innocence, as the Secretary of State must do when refusing an application for compensation under the amended section 133 on the ground that the fresh evidence does not demonstrate the applicant’s innocence. Furthermore, the implication of para 128 of the European court’s judgment in Allen - a category 3 case - is that even in cases where there has not been an acquittal “on the merits” in that sense, as may be the position in the present cases, it is nevertheless impermissible for the criteria for awarding compensation to “[call] into question the innocence of an acquitted person or to require any assessment of the applicant’s criminal guilt”. If the appellants’ criminal guilt is to be assessed, they are entitled under the Convention to the protections afforded in criminal proceedings, including the benefit of the presumption of innocence.

So far as the European court’s comments about Lord Steyn’s speech are concerned, the court appears to me to have understood that Lord Steyn required the applicant’s innocence to be established by a new or newly discovered fact. Its comments seem to me to provide some support for my conclusion. The critical question does not however turn on how the court’s references to Lord Steyn’s speech are to be construed, but on how the approach to article 6(2) laid down by the court applies to section 133 in its amended form. For the reasons I have explained, the criterion laid down in section 133(1ZA) is in my opinion incompatible with article 6(2).

Counsel for the Secretary of State submitted, however, that a violation of article 6(2) was avoided by means of the Secretary of State’s statement, in each of the decision letters, that nothing in the letter was intended to undermine, qualify or cast doubt upon the decision to quash the conviction, and that the applicant was presumed to be and remained innocent of the charge brought against him. I am unable to agree that this statement ensures that article 6(2) is respected. The application of a test which in substance infringes the presumption of innocence is not rendered acceptable by the addition of words intended to avoid a conflict with article 6(2), if the overall effect is nevertheless to undermine a previous acquittal. The point is illustrated by the case of Hammern v Norway, where the operation of a statutory test which required the applicant to prove that he did not perpetrate the acts forming the basis of the charges was incompatible with article 6(2), notwithstanding a statement in the decision that “I should like to stress that the refusal of a compensation claim does not entail that the previous acquittal is undermined or that the acquittal is open to doubt”. The European court commented at para 48 that it was “not convinced that, even if presented together with such a cautionary statement, the impugned affirmations were not capable of calling into doubt the correctness of the applicant’s acquittal, in a manner incompatible with the presumption of innocence”. That comment is equally apposite in the present case.

Finally on this issue, counsel for the Secretary of State submitted that, in order for this court to find that section 133(1ZA) was incompatible with article 6(2), it would have to go significantly further than did the European court in Allen, contrary to the principle expressed in R (Ullah) v Special Adjudicator[2004] UKHL 26; [2004] 2 AC 323, para 23. That argument cannot be accepted. The conclusion which I have reached is based on principles which were already well-established before the case of Allen, and which received the approval of the Grand Chamber in that judgment.

For these reasons, I conclude that the definition of a “miscarriage of justice” introduced by section 133(1ZA) of the 1988 Act is incompatible with article 6(2) of the Convention, and would have made a declaration to that effect.

I agree with Lord Reed that the appeals in these cases should be allowed and that the declaration of incompatibility which he proposes should be made.

It is important to keep clearly in mind that the focus of the case is on the compatibility of section 133(1ZA) of the 1988 Act with article 6(2) of ECHR. The starting point for any discussion of this question must be whether the article is engaged by decisions taken under section 133. For the reasons so compellingly given by Lord Reed, such decisions do fall within the ambit of article 6(2). Inasmuch as the decision in Adams suggested otherwise, it should not be followed. In any event, as Lord Reed has demonstrated, the decision in that case conflated the questions whether article 6(2) was engaged and whether it had been breached.

Lady Hale agrees that article 6(2) is engaged - see para 77 of her judgment. Lord Mance in paras 35-53 of his judgment discusses whether article 6(2) should be “applied” to decisions taken under section 133. As he has pointed out, recent case law from the Strasbourg court has focused on the question whether there is a sufficient link between the impugned decision and the second aspect of the article 6(2) obligation. But, on Lord Mance’s analysis, the focus is not concerned with the question whether the article was engaged but rather on whether it has been violated. I do not construe his judgment, therefore, as suggesting that this species of decision lies outside the ambit of article 6(2).

Lord Wilson agrees (albeit with reluctance) with Lord Reed, that, if article 6(2) has the meaning ascribed to it by the ECtHR, in particular in the Allen case, section 133(1ZA) of the 1988 Act is incompatible with it. Although he declines to follow the case law of Strasbourg on the question of the meaning of article 6(2), I detect nothing in his judgment which suggests that he would find that decisions made under section 133 did not fall within its ambit, if interpreted in accordance with that case law.

Lord Hughes has said that article 6(2), in its second aspect, applies and thus governs subsequent proceedings when there is a link between them and the previously concluded criminal proceedings. In contrast to Lord Mance, it would appear that Lord Hughes considers that the existence of a link was prerequisite to the engagement of article 6(2). But, Lord Hughes’ judgment does not appear to me to be inconsistent with acceptance that the link is present where a decision under section 133 requires to be taken.

At para 99(c) of his judgment Lord Hughes sets out four considerations said to be indicative of the likelihood of the existence of a link, all of which, apart possibly from the final one, seem to be present in this case. They are present where: (i) an analysis of the criminal judgment must be undertaken; (ii) where a review or evaluation of the evidence in the criminal file must take place; (iii) where there has to be an assessment of the applicant’s participation in some or all of the events leading to the criminal charge; and (iv) where comment must be made on the subsisting indications of the applicant’s possible guilt.

Plainly, scrutiny of the criminal judgment must underpin any decision under section 133; likewise, a review of the evidence against an applicant is indispensable; and this must include an assessment of his participation in the events which led to the criminal charge. The only possible debate is as to whether “comment … on subsisting indications of the applicant’s possible guilt” requires that a statement be made by the decision-maker or merely that a judgment be reached by him on these questions: does contemporaneous information lead to the conclusion that the applicant has been fully exonerated; or that he could never have been properly convicted; or whether sufficient new material has been adduced which rendered the conviction unsafe on the basis that a jury might or might not have convicted him had such material been produced at his trial. It seems to me that the decision under section 133 will inevitably require a judgment to be made on those issues and, if that is what is required to meet Lord Hughes’ final criterion, the decision plainly comes within the ambit of article 6(2).

Lord Lloyd-Jones does not directly address the question of the engagement of article 6(2) as opposed to its possible violation but, as with Lord Wilson’s judgment, I detect nothing in his judgment which is counter indicative of acceptance that article 6(2) is at least engaged by decisions made under section 133.

In light of all this, it appears to me that there is general agreement among the members of the court - or, at least, no overt dissent, that decisions made under section 133 fall within the ambit of article 6(2). The question to be concentrated upon, therefore, is whether the context set by section 133(1ZA) involves an inevitable conflict with the article. Put more simply, if a decision as to whether a person whose conviction has been quashed is to receive compensation only if he shows that he was innocent, is such a requirement compatible with article 6(2)?

There has been much erudite discussion in the judgments of other members of the court about the nature of innocence and the inaptness of the criminal trial to investigate and pronounce upon the question whether a defendant is innocent, as opposed to not being proved to be guilty. I do not propose to add to that discussion beyond observing that, inevitably, there will be many who are charged with or tried on criminal offences who are truly innocent but are unable to establish their innocence as a positive fact. That undeniable circumstance must form part of the backdrop to the proper approach to the application of article 6(2) of ECHR.

It seems to me that much of the jurisprudence on the “second aspect” of the sub-article has been influenced, albeit perhaps not explicitly, by the dilemma that this presents. The opportunity to proclaim one’s innocence and the right to benefit from the recognition and acceptance of that condition lies at the heart of much of the dispute in this case and much of the case law of the Strasbourg court on the subject. But an inevitable sub-text is that establishing innocence as a positive fact can be an impossible task. This is especially so if conventional court proceedings do not provide the occasion to address, much less resolve, the issue.

On the other hand, those who have been acquitted simply because the properly high standard for criminal conviction has not been met, but against whom real suspicions as to guilt remain, should not be able to shelter behind the shield of innocence that article 6(2) establishes. In particular, they should not be immune from civil suit from their victims when a less onerous burden of proof as to their involvement in the activity alleged in the criminal proceedings is involved.

It would be idle for me to recapitulate on the extensive examination of the case law of ECtHR that has been undertaken by the other members of the court. I consider that Lord Reed has convincingly demonstrated (in paras 161-175 of his judgment) that there is a “clear and constant” line of jurisprudence from that court which establishes that the relevant question is “whether there was a link between the concluded criminal proceedings and the compensation proceedings, having regard to the relevant considerations” set out in para 104 of the judgment in Allen. For the reasons that Lord Reed has given, I consider that such a link is clearly established. The “relevant considerations” in this context will, of course, include the circumstances of the applicant’s ultimate acquittal of the charge against him. If this is on the basis of a doubt as to whether he should have been acquitted, he will not be able to avail of the article 6(2) protection; if, on the other hand, he can show that he ought never to have been charged or convicted, he will.

I do not agree with Lord Mance’s proposition that “the real test is, or should be, whether the court in addressing the civil claim has suggested that the criminal proceedings should have been determined differently” (para 47 of his judgment). There are two fundamental objections to that formulation of the test. The first is that it would cut out a swathe of deserving applicants when they have not been able to prove that they are innocent when they are in fact. The second is that their fate is determined on the phraseology which happened to be chosen by the court.

For these reasons and those much more fully expressed by Lord Reed, I would make the declaration of incompatibility which the appellants seek.