Justice for the blackest malefactors? Determinate prison sentences, early release, and the ECHR

Abstract

In the light of the High Court's decision in R (Khan) v Secretary of State for Justice [2020] 1 WLR 3932 this paper contends that a revised approach to the interpretation of Articles 5 and 7 of the European Convention on Human Rights is needed. The paper argues that the Article 5 ECHR right to liberty and security plays a developing, though overlooked, role in the context of regulating determinate prison sentences. English law's conclusion that Article 5 of the ECHR has little to offer in this context is wrong and needs to be reconsidered. Equally, a more generous interpretation of Article 7 of the ECHR is now required: an approach which reflects the reality of determinate sentences.

Introduction

In Khan counsel for the Secretary of State contended that ‘it is well established in the case law that there is no right to early release … [it is] always open to Parliament to change the early release provisions’ relating to serving prisoners.¹ Sir James Eadie QC also contended that ‘a change to the administration of a penalty, by an alteration to the early release provisions or the like, will not engage Article 7’.² The High Court agreed, concluding that, amongst other things, that neither Article 5³ nor Article 7⁴ of the ECHR had any application with respect to retrospective changes to early release. This conclusion is wrong. Indeed, the court missed an opportunity to examine the application of both articles more fully in this important context. This paper will argue first that Article 5 does in fact apply in this context in a subtle and complex manner. This paper outlines how the right to liberty and security enjoys an increasingly important, if somewhat overlooked, relevance in the modern penal context. This paper also argues, secondly, that a more generous interpretation of ‘penalty’ contained in Article 7 is possible in preference the literal one adopted by domestic and European courts, and that consequentially changes to early release fall within the scope of this important Convention right.

1. Background

Late on the afternoon of 29 November 2019 Usman Khan, who was attending a conference on the rehabilitation of offenders organised by the Cambridge Institute of Criminology in Fishmongers’ Hall in London, began stabbing other attendees. Before he was shot dead by police, Khan murdered two people and wounded three others. Khan had previously been convicted of engaging in conduct in preparation for acts of terrorism contrary to section 5(1) of the Terrorism Act 2006 and received an indeterminate sentence of 16 years’ imprisonment, with a minimum term of eight years before he was eligible for early release.⁵ Khan was subsequently released in December 2018 with extensive licence conditions. Although the conditions of Khan's early release prevented him from entering London, those conditions were varied so he might attend the conference. A few weeks later, on 2 February 2020, another prisoner released on licence, Sudesh Amman, stabbed two members of the public on Streatham High Road. Amman was also shot and killed by armed police. In November 2018 Amman had been convicted of possessing a document containing terrorist information, and seven counts of disseminating a terrorist publication. Amman was sentenced to three years and four months’ imprisonment.⁶

As a result of these attacks the government announced that it would immediately introduce emergency changes to sentencing law. The changes would end the automatic early release of terrorist offenders at the halfway point of qualifying custodial sentences. In future the government intended that such offenders would only become eligible for early release after two-thirds of their sentence had passed, with any decision on release being at the discretion of the Parole Board. Furthermore, the Secretary of State for Justice argued, that as ‘we face an unprecedented situation of severe gravity … this legislation will therefore also apply to serving prisoners’.⁷ There were, unsurprisingly, immediate concerns over the retrospective nature of the proposed changes to sentencing.⁸ Nevertheless, the Terrorist Offenders (Restriction of Early Release) Bill was subsequently introduced into the House of Commons on 11 February 2020, and passed all the Commons stages a day later without a vote. The Bill passed all its stages in the Lords on 24 February and received Royal Assent shortly thereafter. In the House of Lords the Advocate-General for Scotland, Lord Keen of Elie QC, repeatedly maintained⁹ the view that the Bill was compatible with the ECHR: the changes were to the administration of the sentence and not the sentences themselves.¹⁰ Lord Keen's opinion was shared by many, including Lord Pannick QC,¹¹ although other distinguished lawyers disagreed.¹²

(a) Terrorist Offenders (Restriction of Early Release) Act 2020

On one level TORERA 2020 is a simple statute. The Act makes retrospective changes to the early release regime of terrorist offenders who are subject to determinate sentences.¹³ TORERA 2020 inserts into the Criminal Justice Act 2003 a new section, section 247A. The effect of section 247A is to end the automatic early release of terrorist offenders after they have served half of their determinative custodial sentences. Yet the effect of this section is in practice limited. Since April 2019 terrorist offenders will receive a sentence for ‘offenders of particular concern’.¹⁴ The changes in TORERA 2020 are thus limited to a small class of offenders, numbering around fifty. Instead, this class of terrorist offenders will become eligible for discretionary early release under section 247A when they have completed two-thirds of their custodial sentence. The decision to release an offender will be made by the Parole Board and is no longer automatic.¹⁵

(b) Other changes to automatic release

Further changes were subsequently made to other sentences following the Conservative and Unionist Party's Westminster election manifesto commitment to ‘end automatic halfway release from prison for serious crimes’.¹⁶ By statutory instrument the Secretary of State for Justice modified the application of sections 244 and 264 of the Criminal Justice Act 2003 with respect to prisoners serving a fixed-term sentence of seven years or more for a relevant violent or sexual offence.¹⁷ These prisoners were to be released after the two-thirds their sentences had been served. Crucially, in contrast to TORERA 2020, the changes were to have prospective effect. Then in September 2020 the Johnson administration proposed similar changes in its White Paper A Smarter Approach to Sentencing.¹⁸ Crucially, the White Paper proposed that a new power be enacted that would enable the Secretary of State to prevent the automatic early release of certain offenders of significant concern.¹⁹ Thus, any prisoner serving a determinate sentence for a non-terrorist related offence but who in the Secretary of State's opinion posed a terrorist threat or was otherwise a significant danger to the public could be subject to a new executive power. The new power would enable the Secretary of State to prevent the automatic release of the prisoner and instead refer to the Parole Board the question of whether it would be safe to release the prisoner before the completion of their full sentence. These proposals were subsequently enshrined in the Police, Crime, Sentencing and Courts Act 2022,²⁰ which inserted sections 244ZA–244ZC into the Criminal Justice Act 2003.

2. The broader political and constitutional context

Terrorist episodes often create ‘security panics’.²¹ The events detailed above are an excellent example, sharing many of the characteristics and flaws of earlier similar episodes. When we think of terrorism, particularly in the immediate aftermath of such horrendous crimes, we are gripped by a fear of violence and death. But crucially we neglect to consider the likelihood of being involved in it.²² The vast majority of us simply have no idea how probable it is that we might be involved in a terrorist incident. In these circumstances we fall back on our own perceptions. But such events are mercifully very rare, and in practice terrorism kills few people. For instance, a person is much more likely to be killed in a road traffic accident.²³ Yet because of its psychological availability we perceive terrorism as a much greater threat. Consequently, we allow our fear to drive out qualitative judgments, and to justify extraordinary measures.²⁴ Consequently, legislative responses to terrorism thus carry an attendant risk of probability neglect.²⁵ In the aftermath of terrorist outrages a climate of fear and anger often arises.²⁶ In such a climate it is hardly surprising that our representative institutions will neglect probability too. Terrorist attacks, as Gearty notes, demand that ‘something must be done’.²⁷ In fact, ‘a good deal of legislation and regulation can be explained partly by reference to probability neglect when emotions are running high’.²⁸ Almost without fail every terrorist incident or campaign bequeaths a draconian legislative response.²⁹ Unsurprisingly, one of the policy drivers behind TORERA 2020 was the fear of further terrorist offenders re-offending. Yet terrorist offenders have low rates of recidivism, around 3%. Indeed, it might be argued that TORERA 2020 is more likely to increase recidivism.³⁰ Furthermore, terrorist violence and activity is almost always analysed and solved entirely in the form of additional powers and temporary derogations from fundamental rights which over time become permanent.³¹ Down the years these legislative responses to security panics have exercised a ratchet effect.³² During the expansion phase of a security panic, an episode, or credible threats of such an episode, triggers emergency legislation by the executive-dominated legislature, passed without full and proper deliberation. TORERA 2020 was effectively passed in two days.³³ Such hyper legislative activity is deemed necessary to deal with a threat classed as ‘unprecedented’. Inevitably the government seeks powers which it is argued would have prevented the outrage. But such assertions are never tested, largely because the rapid pace of enactment leaves no time, and the nature of the context means much will be taken on trust.³⁴ Yet each episode resets the baseline normalising the previously exceptional and raising society's tolerance of emergency legislation.³⁵ At such times convention rights should enjoy a particular and significant prominence, acting as a bulwark against oppression.

In addition, the enactment of TORERA 2020 must be set against the populist backdrop of contemporary British politics. Penal populism has long been a pernicious feature of the landscape of English criminal justice.³⁶ But more recently this penal populism has evolved through the criminalisation of terror-related offences. Populism, as Lacey rightly contends, has been instrumental to the significant expansion of the state's power to deal with terror-related activities.³⁷ Crucially, counter-terrorist law ‘deploys the very notion of an enemy alien on which populism feeds’.³⁸ To put it another way, terrorist offenders are the ‘devils’ of security panics. However, it might be said that the law represents something more pernicious, namely that is there is ‘something additionally and intrinsically wrong about being a certain kind of person, engaged in a certain kind of activity – an aggravation of blameworthiness that justifies a special criminalisation regime’.³⁹ Yet populism also presents a further challenge in this context. Legal constitutionalism, even in the weak form practised under the Human Rights Act 1998, requires the judicial protection of ‘discrete and insular minorities’.⁴⁰ Indeed, ‘one of the principal reasons for having constitutional rights is that the ordinary majoritarian political process cannot necessarily be relied on to protect minorities’.⁴¹ In the case of prisoners convicted of terrorist offences the situation is arguably at its most acute. Not only are they effectively banished from society for their crimes, they have no voice in the democratic processes which affect them. Prisoners are isolated physically, socially, and politically. And only through the courts will they be able to access a forum for redress. Unsurprisingly, when this happens the courts also find themselves on the receiving end of populist outrage.⁴²

3. R (Khan) v Secretary of State for Justice

Be that as it may, the changes enacted by TORERA 2020 were subsequently challenged by way of judicial review in Khan.⁴³ The applicant was serving a determinate sentence of four years and six months following his conviction at Newcastle Crown Court for several terrorist offences.⁴⁴ Khan was due to be released on licence on 2 March 2020. TORERA 2020 prevented that. In his judicial review application, counsel for Khan claimed that section 247A of the CJA 2003 (as amended by TORERA 2020) was incompatible with Articles 5 and 7 when read with Article 14 of the ECHR. The applicant claimed declaratory relief to that end. The High Court dismissed Khan's application. In its judgment in Khan the High Court examined the compatibility of TORERA 2020 with Articles 5, 7 and 14. The application of Article 14 in this context raises a series of complex questions, and space precludes a full consideration of them. Instead, the focus here is on the arguments surrounding Articles 5 and 7, which arguably are of greater importance. Article 5 received less attention than Article 7 in the judgment of Garnham J. Nevertheless, both Articles 5 and 7 require careful examination.

(a) The Article 5 challenge

To begin with Article 5, Garnham J held that the changes in TORERA 2020 did not affect the sentence of the Crown Court. The continued detention of Khan remained lawful thanks to the authority of the head sentence.⁴⁵ Furthermore, Article 5 did not guarantee any right of early release for prisoners.⁴⁶ Garnham J then considered a line of authority which dealt with the recall of prisoners released on licence.⁴⁷ Indeed, His Lordship concluded that Article 5 added little that had not already been considered under Article 7.⁴⁸ ‘The fact that a prisoner may expect to be released on licence before the end of the sentence’, Garnham J concluded, ‘does not affect the analysis that the original sentence provides legal authority for detention throughout the term’.⁴⁹ Here the court fell into error. As we shall see below, the application of both Convention articles is arguably more nuanced in this context.

(b) The nature of determinate dentencing

Before examining the application of Articles 5 and 7 it is important to understand the context in which they operate.⁵⁰ During the second half of the twentieth century the nature of prison sentences changed. These changes have important implications in this context. Traditionally, in English law a person sentenced to imprisonment could not expect to be released until the completion of that sentence.⁵¹ Sentenced by Her Majesty's courts, prisoners served their sentences in Her Majesty's prisons. Although prisoners were committed to the custody of the executive for the duration of their sentences, administrative interference with the sentence was rare. In exceptional cases the Home Secretary might exercise the Royal prerogative of mercy.⁵² But it was not until 1967 that there was a formal system of parole. The 1967 parole scheme allowed prisoners to apply to the Home Secretary for release on licence before the end of their sentence. The parole system was not part of the sentencing process; it was a discretionary system of release where the Home Secretary had the final say. The Criminal Justice Act 1967 failed to specify the objectives of the parole, but the system enshrined in law was an administrative one that blended rehabilitative objectives with institutional management.⁵³ Parole was a privilege that prisoners earned rather than a legal right automatically granted or regulated by statute.⁵⁴ Parole was thus a discretionary executive scheme that was not an inherent part of the sentence handed down by the sentencing court. Prisoners had no expectation of parole, domestically at least.⁵⁵

However, in 1991 parole was replaced with a broader scheme of early release. Within that scheme, determinate sentences are imposed on offenders that are not of particular concern.⁵⁶ An offender sentenced to a determinate sentence will be released after serving 50% of the sentence. His release at the 50% point is automatic: it does not depend on the exercise of discretionary powers; it occurs by the operation of law. The position in law, as Lady Hale DP noted in Whitson, is ‘rather stronger than an expectation of release on licence. The prisoner is legally entitled to be released at a certain point in his sentence. This is irrespective of the risk that those responsible for his imprisonment may consider that he poses to the public’.⁵⁷ Thus, as Lady Hale went on to observed in Whitson, ‘in a very real sense … the sentence imposed by the court as punishment for the offence is half the actual term pronounced by the judge (and indeed the judge has to explain this to him when imposing it)’.⁵⁸

At the halfway point determinate prisoners are automatically released. Indeed, early release for determinate sentences of no particular concern does not involve the Secretary of State or the Parole Board. There is no administration of the sentence in the fashion of indeterminate and life prisoners. After his release, the offender is on licence for the remainder of the sentence, subject to whatever conditions the Secretary of State attaches to the licence and where these are breached to the possibility of recall to prison.⁵⁹ Early release, of course, has important penological purposes, as Lord Bingham observed in West:

All, or almost all, determinate sentence prisoners are expected to return to the community on release from prison after serving their sentences. It is in the interests of society that they should, after release, live law-abiding, orderly and useful lives. For a host of practical, psychological and social reasons, the process of transition from custody to freedom is often very difficult for the prisoner. It is accordingly very desirable that the process of transition should be professionally supervised, to maximise the chances of the ex-prisoner's successful reintegration into the community and minimise the chances of his relapse into criminal activity.⁶⁰

The sentencing judge will be fully aware that the offender will be automatically released when passing sentence.⁶¹ Indeed, the judge is obliged to explain the way that the headline sentence will operate to the offender. This explanation is important for, as Lord Steyn explained, ‘a convicted criminal is entitled to know where he stands so far as his punishment is concerned … he is entitled to legal certainty about his punishment’.⁶² Unsurprisingly, both at common law and under statute, a lawful sentence pronounced by a court may not be retrospectively increased.⁶³

The sentence is thus a composite one, ⁶⁴ custody followed by conditional and supervised liberty in the community: ‘The sentence passed is not … a simple statement of the period the defendant must spend in prison … the sentence is in reality a composite package, the legal implications of which are in large measure governed by the sentence passed’.⁶⁵ This composite nature represents the changing nature of the sentence with the passage of time. Or to put it another way, the reconciliation of competing interests in penal policy. In a modern European state, a prison sentence is no longer simply a period of detention. At the beginning of the sentence a prisoner is imprisoned for reasons such as punishment and deterrence.⁶⁶ But the public interest justifications for the custodial detention of the offender have a half-life which decays with the passage of the sentence. At the halfway point, society's and indeed the individual's interest in rehabilitation and reform⁶⁷ come to the fore through the operation of law. Therefore, early release grants the prisoner a conditional liberty which is qualitatively different to imprisonment, with the intention of ensuring their rehabilitation.⁶⁸ A determinative sentence is not in either law or practice a sentence of imprisonment for the length indicated, save in cases where the offender's behaviour on conditional release has demonstrated that custody is the only appropriate place for them. The nature of English determinative sentences was clearly noted by the Irish Supreme Court in Sweeney v Governor of Loughan House Open Centre.⁶⁹ Sweeney was sentenced to a determinate sentence of 16 years in England but was transferred to Ireland at his own request to serve his sentence.⁷⁰ Had he remained in an English prison he would have been granted early release after eight years by operation of law. But in Ireland he remained in prison beyond that point, which he contended was an unlawful deprivation of his liberty. The Irish Supreme Court considered the nature and duration of his sentence and concluded that the sentence as an unqualified one of 16 years would fly in the face of the reality of the situation:

whatever descriptive terms are applied to the sentence imposed on the appellant by the English court, objectively it consisted of one period of imprisonment or ‘deprivation of liberty’, to use the language used in the Act in defining a ‘sentence’, and a second period of liberty under licence which occurred by operation of law and not by virtue of any remission, discretionary or otherwise, granted by the English authorities.⁷¹

Thus, Sweeney's sentence was properly characterised as one of eight years’ imprisonment.⁷² It is high time that English law demonstrated a similar realism.

(c) The evolving nature of the right to liberty: Del Río Prada v Spain

A new realism was to the fore in the judgment of the Grand Chamber of the European Court in Del Río Prada v Spain,⁷³ where the court was careful to distinguish between the de jure position of the applicant and her de facto position. In Del Río Prada, the applicant was imprisoned as a consequence of her eight convictions for terror-related offences. The cumulative sentences handed down at trial amounted to 3,000 years. But in accordance with the provisions of the Criminal Code (1973) the applicant's various sentences were consolidated into a 30-year term of imprisonment. From the head sentence the applicant was told that she would benefit from remission with the effect that she would be released in 2008. However, in 2006 the Spanish Supreme Court revisited its jurisprudence and departed from the governing case law which had been used to calculate the applicant's release date. The court adopted a new doctrine, the so-called ‘Parot doctrine’, which meant that the sentence discounts and remission were applied to individual sentences, and not the 30-year head term. The Parot doctrine applied retrospectively, with the effect that the release date of Del Río Prada became 2017 and not the previously indicated 2008. In her application to the European Court Del Río Prada argued that this retrospective change by the Supreme Court was an infringement of her Article 5 ECHR right. The Grand Chamber agreed. Crucially, the Grand Chamber observed there was a difference between Article 7 and Article 5 of the ECHR. For ‘while Article 7 applies to the “penalty” as imposed by the sentencing court, Article 5 applies to the resulting detention’.⁷⁴ Under Article 5(1)(a) detention must follow and depend upon a conviction.⁷⁵ And with the passage of time the link between the sentence and the deprivation decays.⁷⁶ These are longstanding principles of Convention law. However, the court went on to observe that:

the causal link required under sub-paragraph (a) might eventually be broken if a position were reached in which a decision not to release, or to re-detain a person, was based on grounds that were inconsistent with the objectives of the sentencing court, or on an assessment that was unreasonable in terms of those objectives. Where that was the case a detention that was lawful at the outset would be transformed into a deprivation of liberty that was arbitrary and, hence, incompatible with Article 5.

Although Article 5 of the ECHR does not guarantee the right to early release, the situation may be different where national law contains such an entitlement of a non-discretionary nature.⁷⁷ It is at this point that the execution of the sentence becomes material and can affect the Article 5 ECHR right to liberty and security.⁷⁸ And while the head sentence will in these circumstances provide the causal link required by Article 5(1)(a), the principle of legality inherent in that provision⁷⁹ demands that it be sufficiently foreseeable in its application.⁸⁰ Thus the Grand Chamber concluded ‘compliance with the foreseeability requirement must be examined with regard to the “law” in force at the time of the initial conviction and throughout the subsequent period of detention’.⁸¹ In Del Río Prada, while the applicant's de jure sentence was not varied by the adoption of the Parot doctrine, in contrast her ‘effective detention’ was.⁸² In fact, the loss of remission caused the applicant's period of effective detention to increase by nine years. In these circumstances, the principle of legality required the court to consider whether this change was reasonably foreseeable at the point when the applicant's sentence became final. In other words, was it reasonably foreseeable at the point of conviction that in its application that the sentence would change to authorise the continuing detention of the applicant? In Del Río Prada's case it was not reasonably foreseeable that the methods used to calculate her remission of sentence would change as a consequence of the Supreme Court departing from its existing case law. In effect, Del Río Prada served a longer period of imprisonment.⁸³ And this was incompatible with Article 5(1).

The decision can also be put another way. When sentenced, the principle of finality attaches to the sentence.⁸⁴ Finality is an important part of any system of justice.⁸⁵ In the criminal context where the liberty of the individual is at stake, finality is especially important. Once liability to a criminal charge has been proved, the criminal liability of the defendant is merged in his conviction.⁸⁶ That liability is to be discharged by the punishment imposed by the trial court.⁸⁷ In other words, the charge becomes final with the conviction and sentence.⁸⁸ The punishment imposed thereafter must be reasonably foreseeable. And if in law a prisoner would be entitled to early release, then a retrospective alteration to the prisoner's detriment would be incompatible with Article 5. Here the focus is not on the nature of the de jure sentence, but the de facto or practical consequences of the alteration of the sentence administration.⁸⁹ Both must be foreseeable to be compatible with the principle of legality because both sentence and its administration have consequences for the liberty of the individual. The High Court's analysis of Article 5 of the ECHR in Khan is thus disappointing. Indeed, it is arguable that it fails to adequately take account of the right to liberty in this context. Just as the administration of Del Río Prada's sentence was changed retrospectively to her detriment, Khan's early release date was unforeseeably changed by retrospectively changes to the law governing the administration of his sentence. It matters not one jot in Convention law that the change in England and Wales was caused by the legislature, and not the courts. Legislative changes of the sort enacted by TORERA 2020 undermine the finality of the sentence via its administration.

(d) The legitimate expectation of liberty of prisoners: ECHR

But notwithstanding Del Río Prada, Article 5 of the ECHR will also apply in a further way: one that has been overlooked in the English case law. It is trite to observe that when a prisoner is detained under a prison sentence their right to liberty and security is limited. Obviously, a custodial sentence through its very nature limits that right. But a custodial sentence does not negate the right to liberty and security. In other words, a prisoner continues to enjoy the protection of that Convention right, even if the courts have been uncertain to what extent.⁹⁰ In Clift Lord Bingham concluded that ‘the right to seek early release, where domestic law provides for such a right, is clearly within the ambit of Article 5’.⁹¹ But beyond this dictum the application of the right is not entirely clear, although it is arguably clearer after Del Río Prada. Be that as it may, it is nevertheless possible that Article 5 of the ECHR does apply in this context through a legitimate expectation to early release based on operation of law.⁹² How might that happen?

Legitimate expectations frequently arise under Convention rights. For example, in Scoppola v Italy (No 2) ⁹³ the applicant elected for trial under a summary procedure on the expectation that if convicted he would not receive a life sentence. In doing so the applicant waived his right to a full trial including several due process guarantees. However, before the trial commenced the legislature amended the Code of Criminal Procedure and removed the reduced sentence for those electing to stand trial under the summary procedure, replacing it with a life sentence. The European Court held that when the applicant elected to stand trial under the summary procedure a legitimate expectation arose. The subsequent retrospective changes by the legislature frustrated that legitimate expectation which had arisen when the applicant had relied on the existing law to make decisions relating to his trial: ‘It is contrary to the principle of legal certainty and the protection of the legitimate trust of persons engaged in judicial proceedings for a State to be able to reduce unilaterally the advantages attached to the waiver of certain rights inherent in the concept of fair trial’.⁹⁴

Scoppola concerned the right to a fair trial, but legitimate expectations have also arisen under Article 5 of the ECHR, as Campbell and Fell v United Kingdom ⁹⁵ shows. Campbell and Fell concerned the loss of remission by serving prisoners. In Campbell's case this amounted to 570 days. Before 1991 remission enabled prisoners to shorten their sentences through good behaviour. Remission was technically a privilege and not a legal right. It was thus different to early release under a determinative sentence. Nonetheless, the European Court held that remission created in a prisoner ‘a legitimate expectation that he will recover his liberty before the end of his term of imprisonment’.⁹⁶ And while technically the removal of remission did not amount to a deprivation of liberty, the decision to do so undoubtedly affected the applicant's right to liberty, coming close to a deprivation. Therefore, the court concluded that the disciplinary process for depriving the applicant of remission should be considered a ‘criminal charge’ and attract the protection of Article 6 of the ECHR.⁹⁷ While this is a welcome decision, the reasoning of the court left much to be desired. Though the European Court concluded that the applicant's Article 5 ECHR interests, principally in the form of a legitimate expectation, were affected by the decision to revoke his remission, the basis for the decision was somewhat opaque.

Nonetheless, in the subsequent case of Ezeh and Connors v United Kingdom ⁹⁸ the Grand Chamber noted that the earlier decision in Campbell, namely that the loss of remission ‘came close to, even if it did not technically constitute’ a deprivation of liberty, had needed to be framed in this manner to engage the guarantees of Article 6 as a ‘criminal charge’.⁹⁹ Ezeh concerned the imposition of additional days imposed by way of penalty for offences committed in custody and the absence of legal representation for those facing such proceedings. A clearer deprivation of liberty was thus present in Ezeh than in Campbell. But even so, domestically the English courts had held that the additional days were covered by the original sentence.¹⁰⁰ In Carroll one of the applicants received an additional 21 days for disciplinary matters during his two-year sentence. The effect of the additional days was simply, as Lord Woolf LCJ put it, ‘to postpone the claimants’ release on licence. The awards clearly had a practical effect so far as the claimants were concerned and that practical effect was to postpone their release. But there was no question of their sentence being increased as a matter of law’.¹⁰¹

Of course, the European Court's jurisprudence here represents a broader approach appropriate to the interpretation and application of Convention rights. Whatever the approach of domestic law, the Convention requires an examination of the realities of the situation in which the applicant finds himself. The European Court will not simply look at the appearances.¹⁰² The court will look at the purpose and effect of the measure contended to infringe a Convention right.¹⁰³ In cases involving fundamental rights and freedoms this is axiomatic.¹⁰⁴ In Ezeh the European Court concluded that although domestic law might class the additional days as part of the original sentence, and thus within the scope of the exception in Article 5(1)(a) ECHR, the approach of domestic jurisprudence did not go ‘to the heart of the question’.¹⁰⁵ The fiction that the additional days were covered by the head sentence could not disguise the fact that a new deprivation of liberty that had occurred.¹⁰⁶ In other words, because of the additional days the prisoner would be detained in prison longer than expected, thanks to the intervening disciplinary proceedings which were unconnected to the original conviction and sentence. Khan is, of course, not on all fours with Ezeh. But the Khan's liberty interest under Article 5 is arguably stronger than those in Campbell and Fell. As in Campbell and Fell, Khan enjoyed a legitimate expectation under Article 5 of early release. But this was an expectation which – unlike in Campbell and Fell – was enshrined in law. To argue that changes to the early release regime have no effect on Article 5, as the High Court did, is wrong.

(e) Other Article 5 ECHR rights and principles

This argument is further reinforced by two additional points. First, when interpreting Article 5 of the ECHR it must be kept in mind that the Convention is a ‘living instrument’ which demands ‘increasingly high standards’ in the protection of human rights.¹⁰⁷ Recently the European Court has moved away from interpreting the right to security in Article 5 of the ECHR as purely auxiliary to the right to liberty.¹⁰⁸ But although the right to security has assumed greater prominence in the Convention, it remains wedded to the protection of physical liberty.¹⁰⁹ Nevertheless, in the present context the emphasis that the right to security places on ex post facto authorisations of detention is arguably of assistance. In a series of cases from Russia the European Court noted that any ex post facto authorisation of detention ‘is incompatible with the “right to security of person” as it is necessarily tainted with arbitrariness’.¹¹⁰ The retrospective alteration of a law which has the effect of depriving the individual of their physical liberty is arbitrary, and thus incompatible with the right to security, no less than a retrospective application of an existing law.¹¹¹

Secondly, prisoners under sentence are protected by the principle of rehabilitation inherent in Article 5 of the ECHR.¹¹² This is a nascent but important principle.¹¹³ The European Court has long recognised that the original justification for the deprivation of liberty for an under-sentence prisoner decays with time.¹¹⁴ More recently the court has recognised that the principle of rehabilitation assumes greater prominence as the release date nears.¹¹⁵ This prominence flows from the progression principle where ‘a prisoner should move progressively through the prison system thereby moving from the early days of a sentence, when the emphasis may be on punishment and retribution, to the latter stages, when the emphasis should be on preparation for release’.¹¹⁶ All prisoners, not just those sentenced to life imprisonment, are entitled to know at the beginning of their sentences when they will be released and under what conditions.¹¹⁷ A life sentence which fails to provide these details when it is imposed will be incompatible with Article 3 of the ECHR.¹¹⁸ And arguably any sentence that fails to provide such details is incompatible with Article 5 too, for every prisoner enjoys some expectation of liberty. Rehabilitation, of course, has important purposes for both prisoner and society. Indeed, the European Court has consistently held the rehabilitation of the individual affords protection for their right to dignity.¹¹⁹ In fact, the dignity of the individual is no less at issue here than with whole life sentences¹²⁰ or oral hearings before the Parole Board.¹²¹ Retrospective changes to the sentence which are impossible to influence or change will undoubtedly lead to feelings of unfairness. And as Lord Reed noted, albeit in a different though related context, ‘the potential implications for the prospects of rehabilitation, and ultimately for public safety, are evident’.¹²² Furthermore, in this context Parliament's precipitate decision to alter the early release regime retrospectively and thus undermine the rehabilitation of the affected individuals arguably also engages the right to liberty and security. Article 5 is engaged here because of a coalescence of liberty-related interests protected by the guarantee that a prisoner has a legitimate expectation to liberty which is arbitrarily denied through retrospective changes that result in double punishment extending their period of imprisonment while simultaneously undermining their rehabilitation. Indeed, the intention of the legislature in fixing the date of mandatory release at the half-way point was to provide mitigation in favour of rehabilitation. At the time of sentencing, that is in the circumstances of the case, this is what was considered an appropriate sentence for the offender. Now the legislature has retrospectively altered the effect of sentence without any individualised consideration of the impact on the offender and their rehabilitation. Moreover, the prisoner will now need to demonstrate that he meets the criteria for release once eligible for early release. In other words, automatic release has been transformed into discretionary release with all the challenges that poses for the prisoner. This change represents a significant adjustment to the effective nature of the sentence and the rehabilitation of the prisoner under it.

The principle of rehabilitation has, of course been considered by the UK Supreme Court in the context of those prisoners serving a term of imprisonment for public protection (known as IPP prisoners), most recently in Brown v The Parole Board for Scotland. ¹²³ In Brown the court held that a failure of the government to provide rehabilitative courses to IPP prisoners, necessary so they could demonstrate their suitability for release, would sound in damages in appropriate cases.¹²⁴ In Brown's case he had failed to show any such failing, and his appeal was dismissed. In reaching this conclusion Lord Reed followed the judgment of the European Court in James v United Kingdom.¹²⁵ Under the ECHR, for the imprisonment of offenders to be lawful under Article 5(1)(a) it must be free from arbitrariness. In James the government sought to rely on the risk posed by the applicants to the public as justification for their continuing detention under Article 5(1)(a). But alone, this was insufficient. This legitimate penal aim, the court held, needs to be balanced and reconciled with the duty to rehabilitate offenders, especially in cases where they may pose a risk to the public.¹²⁶ Thus, for the sentence not to become arbitrary and fall outside Article 5(1)(a), prisoners must be provided with reasonable opportunities to address their offending behaviour and any risk they pose. In James the court held that an infringement of Article 5(1) had occurred when the state failed to provide timely access to rehabilitative courses to the applicants, who were IPP prisoners.¹²⁷ In other words, the state had failed to execute the sentence lawfully in Convention terms. This failure would attract just satisfaction for the distress and frustration the situation created.¹²⁸ Similar arguments might have been made in Khan but were not. TORERA 2020's changes disrupted the execution of elements of current sentences in an arbitrary manner. This in turn disrupted not only the prisoners’ expectation of liberty but also their rehabilitation. Whatever rehabilitative progress the prisoners may have made was arbitrarily abandoned by the state because it was assumed – but not shown – that they posed a risk to public safety. Indeed, in no case was there an assessment of the risk the offender posed. The risk was assumed, and their detention arbitrarily continued. Such changes will inevitably lead to the sort of frustration and distress that the applicants experienced in James: changes upon which the European Court found a violation of Article 5(1) and a remedy in damages.

(f) The legitimate expectation of liberty: comparative approaches

It is also significant to note that the European Court is not alone in interpreting the right to liberty in a way that recognises a reasonable expectation of liberty. For example, the US Supreme Court has recognised that prisoners enjoy what it terms a ‘liberty interest’ in parole derived from the Fifth Amendment. A parolee, as the US Supreme Court notes, has a greater liberty interest than a prisoner: ‘the parolee enjoys a liberty incomparably greater than whatever minimal freedom of action he may have retained within prison’.¹²⁹ For instance, a parolee enjoys the liberty to associate with their family and friends, and to work. Of course, a parolee may be subject to much greater governmental control than a citizen at liberty, but their physical liberty is far greater than a prisoner.¹³⁰ Thus the removal of parole can amount to a ‘grievous loss’.¹³¹ While there is no constitutional right to parole, in Greenholtz the US Supreme Court accepted that Nebraska's parole statute created an expectation of release. In other words, prisoners enjoy a liberty interest, which was entitled to a measure of constitutional protection.¹³² Parole played an important instrumental role in rehabilitating the prisoner.¹³³

Similarly, in Canada the Supreme Court has employed the right to liberty in cases where prisoners have contended that their parole interests have been adversely affected. For example, in Gamble ¹³⁴ the appellant argued that she has been incorrectly sentenced at trial to life imprisonment without parole eligibility for 25 years. Gamble was an accomplice to a murder. In the period between the offence and her conviction Parliament changed the law, directing that any such convictions were to be sentenced under the new law and not the law in force at the time of the offence. Gamble contended that instead of a 25-year period of parole ineligibility, the correct sentence meant that she ought instead to be eligible for parole after 10–20 years of her life sentence. Thus, her continued detention in the absence of any consideration of parole was an interference with her constitutionally protected right to liberty,¹³⁵ which was not consistent with the principles of fundamental justice. Gamble did not question her life sentence but instead focused on her access to consideration for parole. The prohibition on parole was, she contended, an unjustified restriction on her right to liberty protected by the Charter of Rights and Freedoms. The Canadian Supreme Court agreed and declared the appellant eligible for parole. Crucially, in giving the judgment of the majority Wilson J held that the deprivation of liberty of the prisoner ought to be reviewed from a qualitative perspective.¹³⁶ The impugned restriction on parole eligibility meant that the appellant would have no chance of parole for the next 13 years, whereas had she been sentenced correctly she would be eligible for parole, although not release. There was clearly a qualitative difference.

Similarly, it is useful here to compare the approach of the English courts with the Canadian. Whereas the right to liberty is indivisible in England and Wales, in Canada the idea of residual or secondary liberty offers a more nuanced and sophisticated approach. Thus in Cunningham v Canada ¹³⁷ the appellant to the Supreme Court had been convicted of manslaughter and sentenced to 12 years’ imprisonment. When sentenced in 1981 Cunningham's parole eligibility would have been fixed at the two-thirds point of his sentence. However, in 1986 the Parole Act was amendment by Parliament with the effect that Cunningham was denied parole in 1989 at the point he had originally expected to be released. Following a detention hearing by the Parole Board under the amended Act, Cunningham was informed that he would be detained until his sentence expired in 1993. Cunningham challenged the statutory changes to the conditions of parole eligibility as inconsistent with his Charter-protected right to liberty. McLachlin J gave the judgment of the Supreme Court dismissing Cunningham's application for habeas corpus. Under section 7 of the Charter Cunningham's liberty rights were undoubtedly engaged. This flowed from the broad notion of liberty which had been adopted at common law following the Miller trilogy¹³⁸ and constitutionalised in Gamble. Cunningham's secondary liberty interest had become engaged in this context by the expectation of discretionary parole.¹³⁹ That expectation did not affect his primary liberty interest, which was governed by the sentence of the court. The appellant was always sentenced to 12 years’ imprisonment and had, both before and after the amendment of the Parole Act, been required as a matter of law to serve the full term. Nonetheless, his secondary liberty interest had been affected by the amendment to the Parole Act. Why? Parole is a status where the prisoner enjoys a ‘better quality of liberty’ under mandatory supervision in the community, rather than in close custody in prison.¹⁴⁰ This, of course, follows from the qualitative approach to the definition of deprivation of liberty taken in Gamble. With good behaviour the appellant could reasonably expect to be released at the two-thirds point on parole. However, the amendments to the Parole Act reduced the possibility of the appellant being granted parole, and with it his expectation of liberty.¹⁴¹ This adjustment in the appellant's expectation amounted to a deprivation of his liberty. McLachlin J then turned to consider whether the deprivation was consistent with the fundamental principles of justice. In other words, whether the changes to parole represented a fair balance.¹⁴² The balance to be struck here lay between on the one hand the interests of society in protecting its members from dangerous offenders and on the other the interests of the prisoner, including his expectation of liberty. Moreover, the changes to the administration of prison sentences and parole were not unique. However, in this case the changes to the Parole Act had provided due process protection for prisoners intended to prevent arbitrary and capricious decision making. Thus, both substantively and procedurally, a fair balance had been struck.¹⁴³

In both the United States and Canada courts have examined the application of the right to liberty in the context of parole and early release. While there is no right to parole or early release in either constitution, the courts in these two jurisdictions recognise that prisoners enjoy a constitutionally protected expectation of liberty. This of course mirrors the tentative approach of the European Court under Article 5 of the ECHR. But all of these decisions cast light on the correct approach to the interpretation of Article 5 of the ECHR in this context, for while Article 5 of the ECHR does not contain a right to early release as such, it is informed by the principle of rehabilitation. Crucially, that rehabilitative process is buttressed by the prisoner's expectation of early release.

4. Retrospective penal changes

(a) General

Much of the debate surrounding the enactment of TORERA 2020 focused on the retrospective changes to under-sentence prisoners. The changes which TORERA 2020 enacted were unusual.¹⁴⁴ At common law there is a presumption that Parliament does not intend to legislate retrospectively.¹⁴⁵ It has long been a principle of the common law that offenders are to be tried and sentenced according to the law in force at the time when their offences were committed.¹⁴⁶ In the penal context statutes are to be interpreted to operate prospectively.¹⁴⁷ An offender ‘is not to be substantially prejudiced by laws construed as having retroactive effect unless Parliament's intention that they should have that effect is plain … the blackest malefactor is as much entitled to the benefit of that presumption as anyone else’.¹⁴⁸ This longstanding principle also ensures ‘that existing prisoners should not be adversely affected by changes to the sentencing regime after their conviction’.¹⁴⁹ On one level the basis of the rule is no more than an application of the constitutional principle of fairness.¹⁵⁰ But at the same time the rule is a specific application of the right to liberty. In short, retrospective changes to an existing sentence result in more time incarcerated. ¹⁵¹ In Khan's case, however, the common law presumption was displaced. TORERA 2020 was clearly intended to have retrospective operational force and effect.

(b) ECHR

In the absence of assistance from the common law, applicants such as Khan must rely on Article 7. Under the terms of Article 7 states must not impose ‘a heavier penalty … than the one that was applicable at the time the criminal offence was committed’. Much therefore turns what constitutes a ‘heavier penalty’ for the purposes of Article 7. Domestically, the question first arose in Flynn,¹⁵² a case which involved a Convention challenge to changes to the early release of life prisoners enacted by Convention Rights (Compliance) (Scotland) Act 2001. The Act operated to the detriment of the appellants by retrospectively altering the expectations of review which they had all been formally given by the Parole Board. A majority of the Privy Council held that Article 7 was engaged, but the law should be read in a way that any incompatibility was removed.¹⁵³ The transitional provisions in the Act would, if left uncorrected, not only offend the principle of neutrality but would also be inconsistent with ordinary notions of fairness.¹⁵⁴ There had been a formal change in the applicable sentence, which ‘was like changing the rules of a race after the runners had set off’.¹⁵⁵ Crucially, Lord Hope noted that European law treatment of the life sentence reflected an evolution in terms of the right to liberty.¹⁵⁶ In short, the reality of life prisoners is such that – in addition to imprisonment – other aspects of the sentence, including the schemes for release and licence, are now to be considered as part of the penalty.¹⁵⁷ The new scheme introduced a new component into the penal system, with the consequence that the prisoners would serve longer in custody.¹⁵⁸ The change would constitute a heavier penalty and would be incompatible with Article 7. In contrast to the majority, however, Lord Carswell defined the ‘applicable penalty’ in a literal way. The penalty for murder, namely life imprisonment, was the same both before and after the 2001 changes.¹⁵⁹ There had been no increase in punishment. Article 7 was not therefore engaged.

When the same question subsequently came before the House of Lords in Uttley ¹⁶⁰ the opinion of Lord Carwell carried the day. Uttley was convicted in 1995 of several historic offences: three rapes and other sexual assaults. Uttley was sentenced to 12 years’ imprisonment. In 2005 Uttley was released at the two-thirds point on licence under the terms and provisions of the Criminal Justice Act 1991. Uttley contended that this was unlawful, and that instead he ought to have benefited from the sentencing regime in force 1991, in other words at the time of his offences. Under the early regime, with his record of good behaviour Uttley would have been released at the two-thirds point, whereupon his sentence would have expired. But under the 1991 Act, thanks to the licence requirements, he was subject to in effect a suspended sentence. Uttley was at liberty on licence subject to recall.¹⁶¹ This, Uttley contended, was a ‘heavier penalty’ for the purposes of Article 7. However, the House of Lords rejected this argument. For Lord Phillips the material question was whether the changes in the sentence imposed a heavier penalty than might have been imposed.¹⁶² The answer was no. At the point of imposition, the maximum sentence was life imprisonment.¹⁶³ This had not changed. Lord Carswell, repeating his earlier opinion from Flynn, held that the penalty was no heavier.¹⁶⁴ The maximum sentence prescribed by law for the offences remained the same. There had been no increase in penalty, and consequently Article 7 did not apply. A change in the remission or parole provisions as an integral part of the sentence was not a change in punishment.¹⁶⁵ Licence conditions were not a discrete penalty.¹⁶⁶ Thus, English law arrived at a literal reading of Article 7. Only an increase of a statutory maximum sentence could, on the authority of Uttley, engage Article 7. However, the error of this interpretation should be immediately apparent. Article 7 has, on this basis, little application to prisoners sentenced under provisions for which the maximum sentence is life imprisonment, there being no heavier penalty in law than a life sentence. In this context the protection afforded by Article 7 ECHR looks theoretical to say the least.

Be that as it may, Uttley then petitioned the European Court, which dismissed his application, confirming the approach of the House of Lords.¹⁶⁷ In the court's judgment the retrospective alteration of licence conditions was not an increase of penalty within the terms of Article 7.¹⁶⁸ In reaching this decision the European Court relied upon the Commission decision in Hogben v United Kingdom where the Commission held that changes to parole were not part of a penalty for the purposes of Article 7.¹⁶⁹ Changes to parole were instead part of the execution or administration of sentences. That was certainly the case in Hogben's application, for at the time of his petition parole was a matter of executive discretion. And, as was noted above, parole was not part of a composite sentence in the way that early release is for prisoners serving determinate sentences.¹⁷⁰ Nevertheless, in Abedin v United Kingdom the European Court confirmed its consistent position that the definition of a penalty did not include early release provisions.¹⁷¹ Here the court distinguished Del Río Prada, concluding that the changes to the early release regime had neither modified nor redefined the penalty.¹⁷² The change to the duration of the applicant's licence conditions was a change in the execution or administration of the sentence, not the sentence itself.¹⁷³ Thus when Khan sought to challenge the changes in TORERA 2020 as incompatible with Article 7, the weight of authority was firmly against him. Nevertheless, counsel for Khan sought to argue that the changes contained in TORERA 2020 were in essence a redefinition of the penalty that fell within the scope of Article 7 in a way comparable to Del Río Prada.¹⁷⁴ However, the High Court was not convinced. There was no redefinition of the scope of the penalty, and no change to the sentence imposed by the trial judge.¹⁷⁵ TORERA 2020 implemented changes to the arrangements for early release, and ‘an amendment by the legislature to the arrangements for early release raises no issue under Article 7 …. a change to those arrangements does not amount to the imposition of a heavier penalty than that applicable at the time the offence was committed’.¹⁷⁶ The High Court dismissed the application. There should be no surprise at the failure of Khan to convince the High Court that Article 7 was engaged by TORERA 2020. But might a more generous interpretation of what constitutes a ‘penalty’ be possible?¹⁷⁷ An interpretation which reflects the realities of a prisoner who expected to be released automatically by operation of law but instead – thanks to retrospective legislation – found himself imprisoned for longer and subject to a new regime for early release.

(c) Retrospective penal changes: Canada

In Canada v Whaling a similar situation arose.¹⁷⁸ Whaling concerned retrospective changes to the Canadian Federal accelerated parole scheme (APR).¹⁷⁹ Whaling, a first-time non-violent offender, was sentenced to a short term of imprisonment. APR, as it stood when Whaling was sentenced, would automatically by operation of law send qualifying offenders to the parole board for consideration against a lower threshold than normally applied. If the parole board had no reasonable grounds for considering that the prisoner would violently reoffend then it had to release the prisoner. Moreover, day parole was triggered earlier, after one sixth of the sentence. In 2011 the Canadian Parliament abolished APR, with the changes applying retrospectively.¹⁸⁰ As a result, Whaling's parole date was changed to his detriment. Before the Canadian Supreme Court Whaling argued that these retrospective changes to the conditions of his sentence were inconsistent with the prohibition against double jeopardy contained in section 11(h) of the Charter.¹⁸¹ At the heart of the appeal was the question of whether these retrospective changes to the parole regime qualified as ‘punishment’ for the purposes of section 11(h).¹⁸² The Canadian Supreme Court unanimously held that they did. Such changes to parole, Wagner J noted, could mean that a prisoner spends more time in penal custody. And imprisonment is ‘the most severe deprivation of liberty known to the common law’.¹⁸³ Thus because of the changes a prisoner will enjoy a longer and more significant deprivation of liberty. Moreover, under-sentence prisoners enjoy a constitutionally protected expectation of liberty derived from the parole regime in force when they are sentenced.¹⁸⁴ A retrospective and punitive change to the conditions of the original sentence will fall within the scope of section 11(h):¹⁸⁵ thus retrospective changes to parole may constitute new punishment. However, changes to that expectation could be justified by the Crown.¹⁸⁶ Some changes to the parole system would amount to a new punishment, but not all. The dominant consideration throughout was ‘the extent to which an offender's settled expectation of liberty has been thwarted by retrospective legislative action’.¹⁸⁷ Crucially, the court held that a retrospective change to parole which increases the period of incarceration before parole, is an additional punishment. The offender had already been tried and sentenced, and the changes increased the punishment in force at the time he was sentenced. In fact, such changes qualify ‘as one of the clearest of cases of a retrospective change that constitutes double punishment in the context of s 11(h)’.¹⁸⁸ Moreover, it is a change that ‘directly results in an extension of the period of incarceration without regard to the offender's individual circumstances and without procedural safeguards’.¹⁸⁹ Turning to the question of proportionality, and specifically whether the limit imposed by the Abolition of Early Parole Act 2011 (AEPA 2011) on section 11(h) was justified, Wagner J noted that the statute was rationally connected to the policy objective of penal reform. However, the provisions of AEPA 2011 did not represent a minimal impairment of the right. Furthermore, the Crown had failed to demonstrate why the prospective application of AEPA 2011 would not meet its legislative objectives, while of course respecting the respondents’ constitutional rights. The changes to parole enacted in AEPA 2011 were disproportionate, and thus unconstitutional.

Whaling is a decision informed by a generous and purposive approach to the guarantee of rights. As such, it more accurately reflects the realities of retrospective changes to schemes of early release. The contrast with domestic and European jurisprudence is stark. The deference to the legislature and executive, disguised through the emphasis on the execution or administration of a sentence falling outside the scope of Article 7, drains that guarantee of any effective content in a context where it is badly needed. It is, therefore, at odds with the European Court's emphasis elsewhere on the principle of rehabilitation.

Conclusion

One of the justifications for TORERA 2020 was public protection. Yet what TORERA 2020 gives with one hand, it takes with the other. In the short term the Act will achieve its purpose, frustrating the early release of several terrorist offenders. If that is seen as in the public interest, it must be remembered that prisons are widely recognised as an important locus for radicalisation.¹⁹⁰ In the longer term the changes in TORERA 2020 are likely to create feelings of unfairness and leave prisoners more susceptible to radicalisation. To put it another way ‘sentences perceived as punitive and unjust contribute to the push factors’ of radicalisation.¹⁹¹ And this is in a context where Muslim prisoners already feel victimised.¹⁹² Moreover, the TORERA 2020 changes also have a dangerous and self-defeating consequence, as Jonathan Hall QC the Independent Reviewer of Terrorist Legislation noted.¹⁹³ Under TORERA 2020 prisoners will only be released when the Parole Board decides it is safe to do so. This means, as the Independent Reviewer of Terrorist Legislation pointedly noted, that those judged to be the most dangerous of terrorist offenders will not be released until their sentence has expired. At this point they will regain their liberty automatically and unconditionally.¹⁹⁴ Thus the changes in TORERA 2020 represent a ‘cliff-edge’ in risk management terms. Whereas under the law as it stood before TORERA 2020 an offender's rehabilitation and reintegration could be effectively managed through licence conditions, these tools have now vanished in the very cases where they will be needed the most. For changes justified based on public protection, this is indeed worrying.

TORERA 2020 was a typical legislative response to a security panic. But already the practice which TORERA 2020 enacted is being more widely adopted. That the ECHR should have little or nothing to offer in this context is surprising. But then, as we have seen, alternative interpretations are possible, and are arguably now essential. In Khan the High Court concluded that Articles 5 and 7 of the ECHR did not apply. It is, however, time to reconsider how these articles both apply in this important context. The right to liberty, for instance, is evolving. Yet none of those evolutions feature in the English jurisprudence. Cases are, of course, confined by their facts. Yet in a series of domestic cases, raising the same or similar questions, our understanding of the right to liberty has not advanced far. But when the jurisprudence of the European Court is carefully examined it is possible to see principles developing that afford a better understanding of the right to liberty and security in this context. Hence, we have observed that the European Court is now prepared to look behind the de jure head sentence and examine the effective duration of a sentence in cases where remission is retrospectively changed to the detriment of a prisoner's right to liberty. Surveying the broader landscape of liberty, we have also seen that the European Court has developed both the principle of rehabilitation and a greater role for the right to security. Equally, we have seen that Article 5 can create a legitimate expectation of liberty. This interpretation is no outlier, reflected as it is in comparable jurisdictions. Taken together these developments represent a richer understanding of the right to liberty, which has significant consequences for the rights of prisoners. Turning to Article 7 of the ECHR, Canadian law shows that a more realistic approach is possible to the interpretation of what constitutes a ‘penalty’ for the purposes of that article. To continue to hold, as English law does, that a penalty within the terms of Article 7 is simply the maximum possible term of imprisonment ignores wider developments in contemporary penal practice, particularly rehabilitation. Ensuring that offenders are effectively rehabilitated is of considerable importance to both the offenders and society. This is particularly the case when dealing with terrorist offenders. Furthermore, the current interpretation of Article 7 is at odds with the increasing emphasis on rehabilitation developing under Article 5. Frequently, we are rightly informed that the ECHR must be read as a whole, and crucially in such a way as to promote consistency and harmony between its guarantees and protections.¹⁹⁵ A more coherent interpretation of the Convention here is thus overdue.