Home | Databases | WorldLII | Search | Feedback Otago Law Review

Ashworth, Andrew --- "Negotiating the fundamental right to personal liberty: four problem cases" [2013] OtaLawRw 1; (2013) 12 Otago LR 1

Last Updated: 23 April 2015

^{F W Guest Memorial Lecture: 14 March 2012}

Negotiating the Fundamental Right to Personal Liberty: Four Problem Cases

Andrew Ashworth*

Personal liberty may be regarded as one of the central values of democratic life, but its boundaries, and thus essential parts of its meaning, remain open to negotiation. Of course it is unreasonable to expect personal liberty to be total, since its extent must be restricted by the principle of equal freedom, which calls for respect for a similar liberty for others. That said, however, there is a sense in which liberty of the person – which is the narrow sense of liberty with which this lecture is concerned – is essential to the realisation of one’s plans, and to the possibility of living life largely according to one’s own preferences. This concept of liberty lies at the heart of most declarations of fundamental rights, and is often phrased in negative terms. The New Zealand Bill of Rights Act 1990 (NZBORA) illustrates this perfectly: the heading for section 22 of the Act is “liberty of the person”, but those words do not appear in the text of the section, which states tersely: “Everyone has the right not to be arbitrarily arrested or detained.” Thus the right to liberty of the person is constituted by freedom from interferences of those two kinds, arrest and detention, and the adverb “arbitrarily” carries most of the freight.¹

If we move to the European Convention on Human Rights, we find a rather different approach to the same subject-matter. Article 5 begins by declaring that “Everyone has the right to liberty and security of person”, but then goes into considerable detail. It states that “No-one shall be deprived of his liberty save in the following cases and in accordance with a procedure prescribed by law,” and enumerates six situations in which it is considered acceptable to deprive a person of liberty. The

^* This is an updated version of the F W Guest Memorial Lecture 2012, delivered at the University of Otago Law Faculty on 14 March 2012, during the author ’s tenure of the New Zealand Law Foundation Distinguished Visiting Fellowship 2012. The author is grateful to the Law Foundation for its support. The author is also grateful to the Arts and Humanities Research Council (UK) for the award of a three year grant to Professors Andrew Ashworth and Lucia Zedner for a study of “Preventive Justice” (ID:AH/H015655/1): some elements of this lecture are discussed in chapter 3 of the ensuing book, A Ashworth and L Zedner Preventive Justice (Oxford University Press, Oxford, 2014). My particular thanks to Lucia Zedner, Patrick Tomlin, Ambrose Lee, and David Mead for comments on a draft of the lecture.

¹ Section 12 of the Constitution of South Africa is in similar terms, as is s 9 of the Canadian Charter of Rights. The International Covenant on Civil and Political Rights, Article 9(1), is somewhat longer but its essence is that “no one shall be subjected to arbitrary arrest or detention.” Thus the word “arbitrary” is crucial here too.

European Court of Human Rights has stated many times that this is an exhaustive list, and that the overall purpose of Article 5 of the Convention is to ensure that people are not deprived of their liberty “arbitrarily,”² although that particular word is not to be found in the lengthy text of Article 5. The Article goes on to deal with the rights of persons who are properly deprived of their liberty, rights which correspond generally with those in section 23 of the NZBORA.

What I want to do here is to explore some of the boundaries of the right to personal liberty. On the one hand, individuals should be able to go about their daily business and to pursue their life-plans without being deprived of their liberty. On the other hand, there is an expectation that the police and other state authorities will take action to preserve public safety and public order, to prevent crimes from occurring and to investigate crimes that have happened. This, in particular the task of preventing crimes, is part of the democratic compromise: citizens may be taken to agree to give up a certain portion of their liberty so that the police can take necessary steps to carry out their preventive purposes. It is for this reason that all human rights instruments recognize exceptions to the right to liberty of the person, designed to allow criminal justice agents such as the police (in circumscribed circumstances) to stop and search suspects, to arrest suspects, and to detain suspects pending trial. The right to liberty of the person is buttressed here by the presumption of innocence, as a protection (again, not absolute) of the citizen from arbitrary official interventions.³ However, essentially, we all have to accept the possibility that the police or other criminal justice agents may lawfully subject us to one or more of these invasive procedures, however innocent we are, if they have reasonable grounds for their suspicions etc. These are significant inroads into the right to liberty of the person, giving rise to a number of contestable situations in which there is room for argument over the application of one or more of the concepts that are to be found in section 22 of the NZBORA – arrest, detention and arbitrariness.

In this lecture I will examine where the boundaries of “deprivation of liberty” should be drawn, and how the tensions and conflicts with the State’s pursuit of prevention should be negotiated. I will not go on to examine the consequential question of the rights of people who have been justifiably deprived of their liberty, important though that topic is. Instead, the focus will be on the prior question, and on four situations:

1. When do police stops amount to deprivations of liberty?

2. To what extent can the police restrict people’s movements for purposes of public safety and public order, a) without depriving

² See eg Engel v Netherlands (1978) 1 EHRR 507 at [58], Ciulla v Italy (1991)

^{13 EHRR 346, and Saadi v United Kingdom (2008) 47 EHRR 427 at [57].}

^{3 See further A Ashworth “Four Threats to the Presumption of Innocence”}

^{(2006) 10 Evidence and Proof 241 and K Ferzan “Preventive Justice and the}

^{Presumption of Innocence” (2013) Criminal Law and Philosophy (Online}

^{First Article).}

^{them of their liberty, or b), if they are so deprived, with justification?}

3. On what grounds can pre-trial detention be justified?

4. Can the preventive detention of ‘dangerous’ people be compatible with the right to liberty?

I shall be drawing mostly on the jurisprudence of the European Court of Human Rights and of the British courts, although occasionally I will summon up the courage to engage with New Zealand decisions. My overall objective, however, is to raise for discussion the questions of principle underlying these four problematic issues.

I When do police stops amount to deprivations of liberty?

An arms fair was being held in the London docklands. The police, using a power to stop and search anyone within a designated area (without the need for reasonable suspicion) under sections 44–45 of the Terrorism Act

2000,⁴ stopped Gillan, who was riding his bicycle in the direction of the arms fair in order to join a demonstration, and also Quinton, who was walking towards the demonstration in her capacity as a journalist and who was taking photographs. Each of them was stopped for around 20 minutes and searched, during which time Quinton was ordered to stop filming the demonstration. Subsequently they brought actions against the police for judicial review, and, when they were unsuccessful, they pursued their applications to the European Court of Human Rights alleging violation of their rights, including their right to liberty. There were several issues in the litigation, but we will focus here on the question whether the police stop amounted to a breach of their rights. The House of Lords held that there was no violation of the right to liberty in Article 5, because persons subjected to a stop and search were, as Lord Bingham put it, merely “detained in the sense of being kept from proceeding or kept waiting.”⁵ In other words, Lord Bingham did not consider the two applicants to have been deprived of their liberty, apparently on a kind of de minimis principle that would exclude stops of short duration, and so there was no need to seek a justification in the exhaustive list of six situations in Article 5(1) of the European Convention.

Lord Bingham also considered whether the stop and search amounted to a breach of Article 8(1) of the Convention, the right to respect for private life, but he held that:⁶

an ordinary superficial search of the person and an opening of bags, of the kind to which passengers uncomplainingly submit at airports, for example, can scarcely be said to reach

⁴ These provisions were abolished by the Protection of Freedoms Act 2012 (see n 24 below), but provisions for suspicionless searches may also be found in s 60 Criminal Justice and Public Order Act 1994 and Schedule

^{7 of the Terrorism Act 2000.}

^{5 R (on the application of Gillan) v Commissioner of Police for the Metropolis}

^{[2006] EWHC 3165; [2006] 2 AC 307.}

^{6 At [28].}

the required level of seriousness. The House of Lords therefore concluded that Article 8 on privacy was not engaged.

When the case was heard in Strasbourg, the European Court of Human Rights adopted a different analysis. The Court focussed on Article 8, and held unanimously that a search of this kind is a prima facie breach of Article 8(1):⁷

Irrespective of whether in any particular case correspondence or diaries or other private documents are discovered and read or other intimate items are revealed in the search, the Court considers that the use of the coercive powers conferred by the legislation to require an individual to submit to a detailed search of his person, his clothing and his personal belongings amounts to a clear interference with the right to respect for private life.

Having reached this conclusion, the Court went on to consider whether this particular stop and search power could be saved by the provisions of Article 8(2) of the European Convention, but held not, on the grounds that the power was too broad and uncertain.⁸

Leaving aside those wider questions, we should look again at two of the points made by Lord Bingham in the House of Lords. First, did these searches involve only a degree of intrusion that fell below the level of seriousness needed to engage the Article 8(1) right? The European Court pointed out that the relevant Code of Practice stated that the police officer “may place his or her hand inside the searched person’s pockets, feel around and inside his or her collars, socks and shoes and search the person’s hair”, as well as turning out the contents of any bag carried by that person.⁹ All these searches take place in public, with the result that items are exposed to the gaze of other people.¹⁰ Although there is a question of degree involved here, the European Court was clear that searches with this amount of intrusiveness do attain the level of seriousness required to engage Article 8(1). Secondly, what about Lord Bingham’s analogy with airport searches? The European Court rejected this on the ground that an air traveller consents to being searched in public, knowing that this is part and parcel of travelling by air, and that an air traveller has the opportunity to abandon items before being searched and indeed can decide not to travel. The stop-and-search powers under the Terrorism Act 2000 (UK) were “qualitatively different” because a person can be stopped anywhere at any time, without notice and without the person having any choice as to whether or not to submit to the search. The coercive and (for individuals) unpredictable application of the statutory powers was therefore thought to set them apart from airport security searches. Yet some may think that the emphasis on consent is

⁷ Gillan and Quinton v United Kingdom (2010) 50 EHRR 1105 at [63].

^{8 For detailed discussion of these aspects of the judgment, see the note at}

^{[2010] Criminal Law Review 416–419.}

^{9 Above, n 7, at [62].}

^{10 In R v Pratt [1995] 1 HRNZ 323 it had been lawful to search the defendant,}

^{but strip-searching him in a public place was held to be unreasonable}

^{and a violation of s 21 of the NZBORA.}

misplaced, not least in a country like New Zealand where there may often be few or no realistic alternatives to air travel for those needing to leave the country. A different approach would be to accept that airport searches do amount to a prima facie violation of the right to respect for private life, but to conclude that, provided there are reasonable rules specifying the purpose and the limits of such searches, such searches are “necessary in a democratic society” so as to ensure public safety.

While most of the European Court’s analysis was focussed on Article 8 and privacy, it did make a brief reference to the implications of the right to personal liberty in Article 5, which had been the central concern of the House of Lords in this case. This is the Article that declares the right to liberty of the person, and so the preliminary issue is always whether a given case involves a deprivation of liberty or a mere restriction on liberty. Lord Bingham decided that this case involved a mere restriction on liberty, since the applicants were kept waiting temporarily, and so Article 5 was not engaged. Without determining this question finally (having decided the case on Article 8 grounds), the European Court indicated its disagreement with that assessment, stating that during the period of 20-30 minutes:¹¹

the applicants were entirely deprived of any freedom of movement. They were obliged to remain where they were and submit to the search and if they had refused they would have been liable to arrest, detention at a police station and criminal charges. This element of coercion is indicative of a deprivation of liberty within the meaning of Art. 5(1).

For the European Court, it was the coercive umbrella sitting over the whole stop-and-search process that led to its classification as a deprivation, rather than a mere restriction, of liberty. This suggests that a relatively brief detention may still amount to a “deprivation of liberty” for this purpose. On this view, Lord Bingham placed his emphasis on the wrong element, making too much of the brief duration and too little of the coercive legislative framework.

The same issue has concerned the Supreme Court of Canada, which in

^{2009 revised its definition of detention for the purposes of the Charter}

^{of Rights and Freedoms. In R v Grant,12 the Supreme Court held that}

^{the key element in the definition of detention is whether the individual}

^{has been “deprived of the right to choose simply to walk away,” and}

^{that in addition to physical restraint the status of detention could stem}

^{from psychological restraint. In this case, G had been spoken to on the}

^{street by a police officer who stood in his path; two other officers then}

^{arrived and stood behind the other officer, blocking G’s way entirely.}

^{The Supreme Court held that the key question was whether a reasonable}

^{person would have concluded that in the circumstances he or she had no}

11. Above, n 7, at 57 referring to Foka v Turkey, App No 28940/95, judgment of 24 June 2008 at [74–79].
12. [2009] 2 SCR 353, per McLachlin C J and Charron J. The stop in this case lasted only 3–4 minutes.

choice but to comply.¹³ This test is less dependent on the police view of the encounter than on the likely perceptions of a “reasonable person:” one judge warned that courts may treat the ‘reasonable person’ as a much more robust and less vulnerable person than the defendant,¹⁴ although in the present case the majority placed considerable weight on the “power imbalance” between one 18 year-old and three police officers and found that G had been detained. This classification of the encounter means that G should have been accorded all the rights of an arrestee, whereas if it had been found that he was voluntarily “helping the police with their enquiries” he would not be treated as having lost his liberty and therefore would not have those rights.¹⁵

In applying the NZBORA the New Zealand courts have likewise emphasised the element of background coercion, and have done so by way of a mixed objective/subjective test: “does the subject have a reasonably held belief, induced by police conduct, that he or she is not free to leave?”¹⁶ This could apply to a whole range of legally permitted stops, with the result that persons lawfully required to provide a breath sample or to submit to an identity check would be “detained” under the terms of s 22, although the courts have excluded brief and relatively insignificant restraints on movement.¹⁷ Where a detention is found, this does not mean that it is necessarily contrary to the NZBORA, but it does mean that it will be open to review for arbitrariness, and that therefore the conduct of the officers involved will be scrutinised closely. Thus in Grayson and Taylor¹⁸ it was held that the proper starting point is to recognise that any search is a significant invasion of individual liberty; the question then is whether it was “unreasonable”, in the terms of s 21 of the 1990 Act. As noted earlier, s 22 of the Act states that arrest and detention will be contrary to the NZBORA if carried out “arbitrarily.” Many of the decided cases concern situations in which the applicable statutory procedure for an arrest has not been followed,¹⁹ but there is authority suggesting that the concept of arbitrariness requires more than a mere departure from a statutory procedure:²⁰

¹³ At [44].

^{14 Per Binnie J, at [176].}

^{15 For discussion, see J Gans, T Henning, J Hunter and K Warner Criminal}

^{Process and Human Rights (The Federation Press, Annandale, 2011) 108–109}

^{and ch 4 generally.}

^{16 R v M [1994] NZHC 1787; [1995] 1 NZLR 242, per Blanchard J at 245; adopted by the Court}

^{of Appeal in Everitt v Attorney-General [2001] NZCA 449; [2002] 1 NZLR 82 and in R v Koops}

^{[2001] NZCA 449; (2002) 19 CRNZ 309. See generally A Butler and P Butler The New Zealand}

^{Bill of Rights Act: A Commentary (LexisNexis, Wellington, 2005) ch 19.}

^{17 Police v Smith and Herewini [1993] NZCA 585; [1994] 2 NZLR 306.}

^{18 [1997] 1 NZLR 399 at 407; to similar effect is the previous statement in}

^{R v Ririnui [1993] NZCA 464; [1994] 2 NZLR 439 that “the intrusiveness and invasion of}

^{privacy involved in any search of the person is such that it ought to be}

^{conducted to no greater extent than the circumstances reasonably require”}

^{(per Hardie Boys J at 442).}

^{19 See, eg, R v Goodwin (No 2) [1993] NZCA 391; [1993] 2 NZLR 390.}

^{20 Neilsen v Attorney General [2001] NZCA 143; [2001] 3 NZLR 433 at 441.}

Whether an arrest or detention is arbitrary turns on the nature and extent of any departure from the substantive and procedural standards involved ... [An arrest is] arbitrary if it is capricious, unreasoned, without reasonable cause: if it is made without reference to an adequate determining principle or without following proper procedures.

This formulation appears to retain a fair amount of judicial flexibility in determining whether the arrest or detention is justifiable and not “arbitrary.”

How, in principle, should it be decided whether a police stop has crossed one of the lines? The first line is whether there is an arrest, in the sense of a deprivation of liberty. As noted above, the Canadian and New Zealand courts have developed a test that depends not on the officer ’s purpose or belief but on the likely perception of a reasonable person in the circumstances of the individual. Thus the police may claim that there was no coercion involved, that the citizen was responding voluntarily to their requests (“helping the police with their inquiries”) and that therefore there was no detention. However, the test of how the police officer ’s actions would appear to the reasonable person recognizes that coercion, and therefore deprivation of liberty, has a psychological element. The purpose of the police is relevant, but the effect of their conduct is central.²¹

This kind of formulation takes seriously the preventive role of the police as well as the liberty of the individual to go about her or his daily business without interruption from the authorities. The test should require the police to consider how their actions may reasonably be viewed by members of the public, and notably by the particular member(s) of the public with whom they are dealing, who may have special needs.

Once it has been decided that the police have crossed the first line, the second line is whether the officer has reasonable grounds, or believes that there are reasonable grounds, to suspect the individual of an offence. Terms such as “reasonable grounds” and “reasonable suspicion” are remarkably vague,²² and it is hardly surprising stop-and-search powers tend to be used disproportionately against certain sections of the community, particularly racial minorities.²³ The new English law, replacing the Terrorism Act provision that featured in the Gillan case, remains unsatisfactory.²⁴ While the new powers are supported by a

²¹ See also R v Inwood [1973] 2 All ER 645.

^{22 In O’Hara v United Kingdom (2002) 34 EHRR 812 the European Court of}

^{Human Rights held that the facts raising suspicion need not be of the level}

^{required to institute a prosecution, let alone to justify conviction. The}

^{Court added that the police may properly act on information received,}

^{if it was honestly thought to furnish reasonable grounds for arrest and}

^{questioning. Compare R v Chehil (2013) 301 CCC (3d) 157 (Canadian SCt).}

^{23 For the British evidence, see Joint Committee on Human Rights}

^{Demonstrating Respect for Rights? A Human Rights Approach to Policing}

^{(HC 320-1) (TSO, London, 2009), and A Sanders, R Young and M Burton}

^{Criminal Justice (4th ed, Oxford University Press, Oxford, 2010) ch 2.}

^{24 Protection of Freedoms Act 2012, inserting a new s 47A into the Terrorism}

^{Act 2000. See E Cape ‘The Counter-Terrorism Provisions of the Protection}

revised Code of Practice, a recent review of policing found that many police officers had developed loose habits in their exercise of stop-and- search powers, with insufficient training and inadequate supervision.²⁵

^{Codes of Practice are important in refining the practical implications of}

^{terms such as “reasonable suspicion,” but they need enforcement.}

The interface between the right to personal liberty and the preventive role of the police is therefore open to considerable negotiation. Since the right to personal liberty is meant to protect individuals against arbitrary loss of liberty, the potential for slippage at the stage of “reasonable suspicion” gives grounds for concern. However, as argued earlier, no democratic country could operate without allowing the police and other authorities to deprive someone of their liberty on reasonable suspicion that they have committed or are about to commit an offence: in furtherance of the State’s preventive role, such an arrestee loses liberty and gains certain rights. Even a wholly innocent person must accept the possibility of arrest on reasonable suspicion, as part of the democratic compromise. These situations are controversial and fraught, however, and ways must be found of providing guidance for both the police and citizens about the contours of their powers and rights. Otherwise, failure to respect these rule-of-law values risks emptying the right to personal liberty of a vital part of its content.

II To what extent can the police restrict people’s movements for the purposes of public safety and public order, a) without depriving them of their liberty, or b), if they are so deprived, with justification?

The last few years have also seen a rather different British challenge to the right to personal liberty in Article 5(1) of the European Convention on Human Rights, in relation to preventive policing. The police have occasionally used the device sometimes known as “kettling” in order to manage potential public disorder, by confining a number of demonstrators within a relatively small area for a few hours in the hope of preventing a volatile situation from turning into major disorder. Thus in Austin and others v Commissioner of Police for the Metropolis²⁶ some 2,000 people were “kettled” or contained in a side street near Oxford Circus, in central London, during a demonstration against capitalism and globalisation. The police stated that they feared a substantial breach of the peace if this action were not taken, and that this justified their actions at common law. The containment was maintained for seven hours; some 400 people were released in the meantime, but some elements of the kettled group were

from Freedoms Act 2012: Preventing Misuse or a Case of Smoke and

^{Mirrors?’ [2013] Criminal Law Review 385.}

^{25 Her Majesty’s Inspectorate of Constabulary Stop and Search Powers: are}

^{the Police using them Effectively and Fairly? (2013) at www.hmic.gov.uk.}

^{26 [2009] UKHL 5, on which see D Mead “Of Kettles, Cordons and Crowd}

^{Control” [2009] European Human Rights Law Review 376, and R Glover}

^{“The Uncertain Blue Line – Police Cordons and the Common Law” [2012]}

^{Criminal Law Review 245.}

violent and the police stated that it was difficult to pursue a programme of phased release. Ms Austin, a peaceful demonstrator, was not allowed to leave; nor were the other 3 claimants, who were uninvolved members of the public who had been taking a break from work. They all sued the police for false imprisonment.

In the House of Lords the leading speech was given by Lord Hope, who identified the purpose of the police’s actions as the crucial element in the case. The starting point is that Article 5 protects the right to liberty of the person, and that Article 5(1) sets out six sets of exceptional circumstances in which a person may properly be deprived of liberty. None of those exceptions applied in this case, but Lord Hope held that in determining whether there has been a “deprivation of liberty” a court must take account of the whole situation. He concluded:²⁷

The importance that must be attached in the context of Article 5 to measures taken in the interests of public safety is indicated by Article 2 of the Convention, as the lives of persons affected by mob violence may be at risk if measures of crowd control cannot be adopted by the police. This is a situation where a fair balance is necessary if these competing rights are to be reconciled with each other. The ambit that is given to Article 5 as to measures of crowd control must, of course, take account of the rights of the individual as well as the rights of the community.

So any steps that are taken must be resorted to in good faith and must be proportionate to the situation which has made the measures necessary. This is essential to preserve the fundamental principle that anything that is done which affects a person’s right to liberty must not be arbitrary. If these requirements are met, it will be proper to conclude that measures of crowd control that are undertaken in the interests of the community will not infringe the article 5 rights of individual members of the crowd whose freedom of movement is restricted by them.

Lord Hope’s reference to Article 2 in the first quoted paragraph will be discussed later. But we should note here that the last sentence of the quotation is problematic, in that he refers to the “interests of the community” as a sufficient justification for the actions taken. Casual references to “the interests of the community” (or, as in the previous paragraph, “the rights of the community”) are inappropriate in the context of Article 5, since there is no reference in the Article to what is “necessary in a democratic society”: what Lord Hope should have said is something along the lines of “the rights of other people who might be affected by the dangerous behaviour that was anticipated.” The argument should not be skewed by any leverage falsely derived from the political irresistibility of some conception of “community safety” or the like.

More will be said about the judgments in the House of Lords as we turn to consider the judgment of the Grand Chamber of the European

²⁷ At [34].

Court of Human Rights in Austin and others v United Kingdom,²⁸ which held by 14 votes to 3 that there was no deprivation of liberty in this case and therefore no violation of the right to liberty of the person guaranteed by Article 5. The Strasbourg majority started from the proposition that the Convention is a “living instrument” which must be interpreted according to the exigencies of the time. This doctrine, which makes sporadic and unpredictable appearances in the Court’s jurisprudence, may have considerable merits; but one of those, surely, is not that it entitles the Court to ignore (rather than to depart from with reasons) its own recent judgments. Thus, as we saw above, in Gillan and Quinton v United Kingdom²⁹ the Court held that a 20-minute police stop did amount to a “deprivation of liberty” within Art 5(1) because “the applicants were entirely deprived of any freedom of movement” and if they had tried to leave the place “they would have been liable to arrest, detention at a police station and criminal charges. This element of coercion is indicative of a deprivation of liberty within the meaning of Art 5(1).”³⁰

For the majority in Austin not even to mention this judgment, let alone to discuss its approach to the issue at hand, is extraordinary and surely not dictated by the “living instrument” doctrine.

In the House of Lords, Lord Hope rested much of his reasoning on the point that the police actions were based on the purpose of preserving public order and ensuring public safety. However, the Grand Chamber recognises that a well-motivated purpose has not hitherto been considered sufficient to determine whether or not a person has been deprived of liberty.³¹ Instead, the majority points to the “definition” of a deprivation of liberty in Guzzardi v Italy³² which refers, among other characteristics, to the “type” and “manner of implementation” of the measure in question. The majority regards this as the solution to the problem in this case: regard should be had to the “specific context and circumstances” of the restrictions, particularly where the restrictions are “unavoidable as a result of circumstances beyond the control of the authorities and are necessary to avert a real risk of serious injury or damage, and are kept to the minimum required for the purpose.”³³

In effect, therefore, the majority is arguing that it is not so much the degree of the restrictions, or the coercive threat that surrounds them, but rather what kinds of justification may be offered for them. This is quite unpersuasive: the issue of whether there is a deprivation of liberty should be distinguished from the issue of whether it is justified. In effect, the majority has smuggled the element of purpose back into its reasoning under the unconvincing guise of the “type and manner of

²⁸ (2012) 55 EHRR 459, on which see N Oreb (2013) 76 Modern Law Review

^{735; the judgment was handed down the day after the F W Guest lecture.}

^{29 (2010) 50 EHRR 1105; above, n 7 and accompanying text.}

^{30 (2012) 55 EHRR 459 at [57].}

^{31 At [58].}

^{32 [1980] ECHR 5; (1980) 3 EHRR 333.}

^{33 Above n 28 at [59].}

^{implementation” of the restrictive measure.}

Moreover, as the dissenting judgment points out, just because a public order rationale is being advanced for particular police action should not mean that that action is treated differently from other deprivations of liberty. The danger is obvious: states could drive a coach and horses through the protections of Article 5 if “public order” were treated as a “trump card” in this way. As the Grand Chamber itself held in the anti- terrorism case of A v United Kingdom, if detention does not fit within the confines of the exceptions enumerated in Art 5(1), “it cannot be made to fit by an appeal to the need to balance the interests of the State against those of the detainee.”³⁴

The majority judgment says surprisingly little about the “conflict of rights” argument adumbrated by Lord Hope. It mentions the point but does not pursue it, save to say:³⁵

Articles 2 and 3 may imply positive obligations on the authorities to take preventive operational measures to protect individuals at risk of serious harm from the criminal acts of other individuals. When considering whether the domestic authorities have complied with such positive obligations, the Court has held that account should be taken of the difficulties involved in policing modern societies, the unpredictability of human conduct and the operational choices which must be made in terms of priorities and resources.

This passage is supported by references to Giuliani and Gaggio v Italy,³⁶ the G8 case from Genoa, and P F and E F v United Kingdom.³⁷ Article 2 (the right to life) and Article 3 (freedom from torture and inhuman or degrading treatment) are probably the most fundamental and most powerful of rights under the Convention, and on one analysis they conflict with Article 5 on the facts of Austin. Tugendhat J found that the police were acting in order to prevent “damage to property, serious physical injury and even death.”³⁸ In the majority’s judgment this becomes diluted to “serious injury or damage”, whereas damage of itself cannot fall within either Art 2 or Art 3. But if it were accepted, pace the majority here, that the “kettling” did deprive the applicants of their liberty, could it not be said that there was a conflict between their Art 5 rights and the rights of other members of the public, other demonstrators and police officers under Arts 2 and 3 not to be subjected

³⁴ (2009) 49 EHRR 625 at [171]; see the earlier analysis of this proposition by D Meyerson “Why Courts Should not Balance Rights” [2007] MelbULawRw 34; (2007) 31 Melbourne University Law Review 873.

^{35 Above, n 28, at [55].}

^{36 (2011) 52 EHRR 112.}

^{37 Judgment of 23 November 2010; in this case the Court held the complaint}

^{inadmissible, stating that the police have a positive obligation to “take}

^{reasonable steps” to protect any Art 3 rights that are at risk, and that}

^{judgments of reasonableness should take account of the operational}

^{difficulties of policing.}

^{38 Above, n 28, at [26].}

^{to serious injury or even death?}

If the majority had taken this argument seriously, then two further points should have been considered. One is that where rights conflict, the proper approach is not to reach for some nebulous “balancing” but rather to pursue the policy of minimum deprivation. Thus, as in Doorson v Netherlands,³⁹ the approach should be to reduce the individual’s right as little as possible, and to provide compensating or counter-balancing measures for any reduction that is thought unavoidable. What this would mean in the present case remains to be worked out. But, secondly, this whole line of argument can only be based on a finding that there were reasonable grounds for the police commander to believe that there was a real risk of serious injury or death. Otherwise the mere assertion of the risk of serious injury becomes the simplest of ways to circumvent the protection of Art 5.

This brings the discussion to a matter touched on hardly at all in the majority’s judgment. The focus of the case was the right to personal liberty of four individuals. Art 5 is one of the more powerful provisions in the Convention – less fundamental than Arts 2 and 3, just considered, but plainly more powerful than many other Articles in the Convention, in the sense that Art 5 contains no override for the public interest, or for what is “necessary in a democratic society” etc. One feature of the right to liberty of the person is that it is personal: it is each person’s right, and no-one should be deprived of that right by proxy, as it were. This was recognised in the less demanding context of Articles 10 (freedom of expression) and 11 (freedom of assembly) by the House of Lords in R (Laporte) v Chief Constable of Gloucestershire⁴⁰ in finding that it was wrong to send a whole coach full of demonstrators (including the applicant) back to London when a few of the passengers were bent on unlawful activity but there was no evidence that the applicant was. The police should start from the point of view of dealing with each person separately, even if operational problems render this difficult. The minority in Austin is surely right, therefore, to call attention to the fact that three of the four applicants were uninvolved members of the public who were coincidentally swept up into the police cordon. This is not to suggest that the right to liberty of a peaceful demonstrator such as Austin is unimportant, but the liberty of the uninvolved people inside the cordon should have been a primary concern of the police and not a small matter easily overwhelmed by “operational necessities.”

This leads on to the question of the compatibility of police tactics with human rights.⁴¹ The Manual of Guidance on Keeping the Peace⁴² sets out

³⁹ (1996) 22 EHRR 630.

^{40 [2006] UKHL 55, per Lord Bingham at [55]; cf Dunlea v Attorney-General}

^{[2000] NZCA 84; [2000] 3 NZLR 136.}

^{41 On which see D Mead The New Law of Peaceful Protest: Rights and Regulation}

^{in the Human Rights Era (Hart Publishing, Oxford, 2010) ch 7 and especially}

^349-356.

^{42 Association of Chief Police Officers (2010) at [2.79] and [2.80], accessed}

three conditions under which the police may resort to “kettling”: good faith on the part of the police, proportionate to the situation making the measure necessary, and enforced for no longer than necessary. This is followed by a list of some 21 court decisions, including Gillan and Quinton v United Kingdom which, as stated above, stands for a very different proposition. However, it seems that the Manual is now broadly compatible with the judgment in Austin v United Kingdom, although “proportionate to the situation” is too loose, and should be replaced by the requirement that the measures taken must be “unavoidable as a result of circumstances beyond the control of the authorities and are necessary to avert a real risk of serious injury or damage, and are kept to the minimum required for the purpose.”⁴³

This would also be consistent with the judgments in R (Moos and McClure) v Commissioner of Police for the Metropolis, arising out of the G20 protests in 2009. At one stage the police kettled a peaceful crowd, for fear that it might imminently be infiltrated by disruptive individuals and thus de-stabilised. The Administrative Court held:⁴⁴

The police may only take such preventive action as a last resort catering for situations just about to descend into violence ... It is only when the police reasonably believe that there is no other means whatsoever to prevent an imminent breach of the peace that they can as a matter of necessity curtail the lawful exercise of their rights by third parties. The test of necessity is met only in truly extreme and exceptional circumstances.

The Court added that the police should also consider less intrusive measures such as cordons or closing off streets before deciding on a course that infringes the right to liberty, thus re-affirming the principle of the least restrictive alternative.⁴⁵ The reversal of this judgment by the Court of Appeal⁴⁶ does not weaken that principle or the propositions in the above quotation.⁴⁷ The Court of Appeal held that the Administrative Court appeared to focus on whether a breach of the peace was imminent at the time, whereas the question in judicial review proceedings should be whether the police reasonably believed that a breach of the peace was imminent. Neither court placed any emphasis on a threat to the right to life in Article 2, as distinct from the potential breach of the peace.

The tactic of kettling may hold advantages for the police, in their preventive role, but, from the point of view of individuals subjected to it, it tends to be applied indiscriminately (so as to detain both peaceful demonstrators and passers-by), it is disruptive of one’s plans to a

at <www.npia.police.uk>; see also Association of Chief Police Officers,

^{The National Decision Model (2012).}

^{43 Austin v United Kingdom (2012) 55 EHRR 459 at [59].}

^{44 [2011] EWHC Admin 857 at [56].}

^{45 For discussion of this principle more generally see A Ashworth and}

^{L Zedner Preventive Justice (Oxford University Press, Oxford, 2014)}

^{ch 3 and 11.}

^{46 [2012] EWCA Civ 12.}

^{47 At [95] specifically endorsing the passage quoted in the text.}

significant degree (because no-one can be sure of being released for several hours), and it can cause inconvenience and discomfort (because of the absence of food, drink and toilet facilities). The arguments adopted by the House of Lords and of the majority of the Grand Chamber in Austin, in holding that there was no deprivation of liberty on the facts, are monumentally unconvincing. They seem to be an example of result- pulled reasoning: because there may in exceptional cases be the need for emergency measures of detention, and this cannot be authorised by any of the enumerated exceptions to Art 5(1) (which are exhaustive), therefore it is necessary to hold that “kettling” does not amount to a deprivation of liberty. If such emergency measures were thought necessary, then these powerful courts should have had the courage to take one of two approaches. First, the courts may have authorised a further exception to Art 5(1), using the “living instrument” doctrine to engage with the reasoning of those judgments which had held that the exceptions to Art 5(1) are exhaustive, and justifying an “emergency measures” exception. Any such exception would have to be authorised only in extreme circumstances, of the kind set out in the passage from Moos and McClure (above); should be limited to the minimum possible duration in respect of each individual; and should be subject to restrictions in order to counter-balance the material deprivations listed at the beginning of this paragraph. A second, alternative approach would be to treat the issue as involving a “conflict of rights”, requiring the court to recognise a conflict with Article 2 and to strive to preserve as much of the Article 5 right as possible or to provide counter-balances to any loss.⁴⁸ In holding that there was no deprivation of liberty involved in several hours of kettling, the courts in Austin have created a position that is both unconvincing and uncertain.

III On what grounds can Pre-Trial Detention be Justified?

This is a question with serious consequences, and yet the way in which I have formulated it pre-supposes that there must be some justification for pre-trial detention, or remand in custody pending trial. Whether there is good reason to take that for granted is the main question to be asked here. Immediately we become aware of the presence of two fundamental rights – the right to liberty and the presumption of innocence. But in this context the former right is usually qualified: thus in s 24(b) of the NZBORA a person who has been charged with an offence has the right to be “released on reasonable terms and conditions unless there is just cause for continued detention”, and the European Convention has a rather convoluted exception in Article 5(3) which is meant to achieve the same end.⁴⁹ Thus, like the loss of liberty involved in arrest on reasonable

⁴⁸ For reasoning of this kind in conflict of rights cases, see Doorson (above, n 39) and Rowe and Davis v United Kingdom [2000] ECHR 91; (2000) 30 EHRR 1.

⁴⁹ “Everyone arrested or detained in accordance with the provisions of paragraph 1 c of this article shall be brought promptly before a judge or other officer authorised by law to exercise judicial power and shall

grounds, it seems that loss of liberty pending trial is a deprivation that must be accepted (in certain circumstances) as part of the democratic compromise. Or is it?

Let us move towards the key issues by explaining, first, what is involved in these decisions. We should mention at the outset that a person arrested at the weekend should still be brought before a court as soon as practicable,⁵⁰ a point that emphasises the need to justify any deprivation of liberty, however short it may appear to the authorities. More significantly, if a case is not disposed of at the first court hearing (because the prosecution or defence or both are not ready to proceed), it will be adjourned until a later date. The question is whether the defendant should be released on bail, either conditionally or unconditionally, or should be remanded in custody (pre-trial detention) for that period. In England, cases that are committed to the Crown Court will often result in a custodial remand, not least because they are the most serious cases. In many cases the period before trial will be lengthy: in 2011 the average period of custody awaiting trial was 9 weeks; in 2009 some 28 per cent of remanded prisoners lost their liberty for between 3 and 6 months, and for a further 8 per cent it was longer than 6 months.⁵¹ We can take it, therefore, that what we are discussing here may be around two months deprivation of the liberty of a person who has not yet been tried for, let alone convicted of, the offence.

What is the purpose of this pre-trial detention? The jurisprudence of the European Court of Human Rights has recognised four purposes as sufficient.⁵² One is the purpose of ensuring that the defendant is present at the trial and does not abscond: thus a significant issue should be the court’s assessment of this defendant’s risk of absconding. This reasoning is closely connected to conceptions of the proper operation of the criminal justice system, which would be impaired if defendants were usually not present at their trial and thus not able to be subjected to a sentence in the event of their conviction.⁵³ A second purpose is the prevention of interference with witnesses or other attempts to interfere with the course of justice: where the court is presented with evidence that such interference is probable, this may justify deprivation of liberty. A third, relatively unusual purpose is the preservation of public order – where the release of a particular person might lead to disorder or to attacks on

be entitled to trial within a reasonable time or to release pending trial. Release may be conditions for guarantees to appear for trial.”

^{50 Whithair v Attorney-General [1996] 2 NZLR 45.}

^{51 Prison Reform Trust Bromley Briefings Prison Factfile (November 2012) at 22}

^{<www.prisonreformtrust.org.uk>; see also E Player, J Roberts, J Jacobson,}

^{M Hough and J Robottom “Remanded in Custody: an Analysis of Recent}

^{Trends in England and Wales” (2010) 49 Howard Journal of Criminal Justice}

^{231 at 237.}

^{52 For elaboration, see B Emmerson, A Ashworth and A Macdonald (eds)}

^{Human Rights and Criminal Justice (3rd ed, Sweet & Maxwell, London,}

^{2012) ch 8.}

^{53 However, trial in absentia remains possible; see ibid, ch 9.}

the person released. In recent years, however, the predominant purpose of custodial remands appears to be a fourth one – the “prevention of crime” or the “risk of committing offences”. In England and Wales this has arisen partly from the fear among victims that, if a person who has (allegedly) subjected them to harm is released on bail, the victimization might be repeated. In part it has arisen from concern about the number of offences being committed by defendants who have been released on bail pending their trial.⁵⁴ The structure of the New Zealand law on bail is similar: s 24 of the NZBORA is supported by s 8 of the Bail Act 2000 (NZ), which identifies three reasons for “just cause” for non-release – a risk of non-appearance in court, of interfering with witnesses or evidence, or of offending while on bail. In New Zealand as elsewhere, care is rightly taken to avoid phrasing this last purpose as “preventing the commission of further offences”, since that might suggest that the defendant is guilty of the offence with which he has been charged but for which he has not yet been tried.

The European Court of Human Rights gives prominence to the presumption of innocence in its judgments on pre-trial detention, but allows the presumption to be overridden. Thus the judge:⁵⁵

having heard the accused himself, must examine all the facts arguing for and against the existence of a genuine requirement of public interest justifying, with due regard to the presumption of innocence, a departure from the rule of respect for the accused’s liberty.

The appearance of the opaque concept of “public interest” in this passage rather detracts from the references to the two principal rights, the right to liberty and the presumption of innocence. However, the Strasbourg jurisprudence has at least insisted on proper evidence and reasoning in relation to the four purposes of pre-trial detention that it has identified. Thus:⁵⁶

the danger of an accused’s absconding cannot be gauged solely on the basis of the sentence risked. As far as the danger of re-offending is concerned, a reference to a person’s antecedents cannot suffice to justify refusing release.

The European Court has also insisted that, before depriving a person of liberty, courts must consider “whether there [is] another way of safeguarding public security and preventing him from committing further offences;”⁵⁷ and the New Zealand Court of Appeal has likewise insisted that “the societal interest must be unable to be met by the

⁵⁴ For fuller discussion and references, see A Ashworth and M Redmayne The Criminal Process (4th ed, Oxford University Press, Oxford, 2010) ch 8.

^{55 Caballero v United Kingdom [2000] ECHR 53; (2000) 30 EHRR 643 at 652.}

^{56 Ibid, citing Yagci and Sargin v Turkey [1995] ECHR 20; (1992) 20 EHRR 505 at [52], on the}

^{first point and Muller v France, judgment of 17 March 1997 at [44], on the}

^{second point.}

^{57 Jablonski v Poland (2003) 36 EHRR 455 at [84].}

granting of bail upon terms as to residence, reporting to police, curfew, non-association, travel restrictions and the like.”⁵⁸ Both courts therefore subscribe to the principle of the least restrictive alternative. Further, in cases where a person is deprived of liberty before trial, the European Court has emphasised the need for “special diligence” in avoiding delay.⁵⁹

Nonetheless, all these worthy principles must be viewed in the light of the Court’s receptiveness to “public interest” arguments, and to balancing various public policy objectives against the two principal rights. So we must return to the question: does the state have a responsibility for the conduct of a person between the time of charge and the time of trial?

Such a responsibility can certainly be asserted in cases where the risk of absconding or risk of interference with the course of justice is relied upon, since these reasons are related directly to the integrity of the criminal trial as a public means of calling the defendant to account.⁶⁰

This does not necessarily establish that detention is justified, but that there is a pragmatic rationale for some kind of coercive restrictions on the defendant during that period, which should be the least restrictive alternative compatible with the relevant purpose. Another necessary step in the process of justification is empirical. If the court has reason to believe that the defendant presents a danger to a particular person or to the public at large, this amounts to a prediction that the defendant would commit offences if released on bail. Yet the grounds for (and accuracy of) such predictions have been little explored. Where the reason for restrictions or detention is the prevention of absconding or the prevention of interference with witnesses, there appear to be no empirical data and the court has to assess the risk on the basis of the evidence adduced (eg previous incidents, threats made). It can be argued that these two justifications are linked to ensuring the integrity of the criminal trial: they should reduce any risk that the defendant will not attend trial, and also reduce the risk that evidence will be tampered with. However, when we turn to the prevention of offending, there is a manifest problem: none of the statistics relating to the risk of persons charged with particular offences committing crimes if allowed bail indicate that offending is more probable than not – even where the person has previous convictions.⁶¹ Such poor prospects of prediction raise serious questions about the justification for relying on this “prevention of crime” purpose; and, in the context of the presumption of innocence, it seems wrong that a probability of around 20–25 per cent should be thought sufficient to justify pre-trial detention. The position may be different if the

⁵⁸ B v Police (No 2) [1999] NZCA 205; [2000] 1 NZLR 31 at 34; see generally Butler and Butler,

^{above, n 16, ch 22.3.}

^{59 Eg Punzelt v Czech Republic (2001) 33 EHRR 1159 at [73]. See also the}

^{decision in Chergui v Police [2006] BCL 363 to the effect that the prospect}

^{of a delay of 2 years before trial is itself a reason for granting bail.}

^{60 See R A Duff, L Farmer, S E Marshall and V Tadros The Trial on Trial,}

^{vol 3: Towards a Normative Theory of the Criminal Trial (Hart Publishing,}

^{Oxford, 2007).}

^{61 For discussion, see Ashworth and Redmayne, above, n 54, at 246–247.}

defendant has previously committed offences while on bail; it may also be different if the defendant is charged with a serious offence and has a previous conviction for a similar level of offence, where the seriousness of the possible harm may be thought to make up for the generally low rate of prediction. ⁶²

If, doubtfully and provisionally, one assumes that these predictions can reach a satisfactory level of accuracy, the next step in the justificatory chain is the assumption that the state has a responsibility to prevent offences being committed by persons who are already formally ‘in the system.’ This crucial assumption has received hardly any analysis. The argument seems to be that, if a person has been charged with an offence, and if the court is unable to deal with the case straight away (which may be because the prosecution or defence or both are not yet ready to proceed), the state has a responsibility to ensure that there is no undue risk of offences being committed by this person during the period before trial. The proposition is straightforwardly protective. The state has a positive obligation to protect the lives and security of people within its boundaries, and a specific obligation to provide protection for individuals whose lives are under known threat.⁶³ The question is whether the state also has a wider responsibility for the conduct of persons who have been charged with an offence but not yet tried, a responsibility that it does not bear for the conduct of other citizens walking the streets. How might such a responsibility be supported? One argument might be that the decision of a public prosecutor to charge a person with an offence tips the balance: such a decision might be regarded as quasi-judicial if taken by a public prosecutor as distinct from a police officer, but, even so, there are difficulties. By the time of the first court hearing the prosecutor may not have complete evidence on all points necessary for the offence charged, and the court is unlikely to be in a position to require this. So the first and other early remand hearings proceed inevitably on a lesser standard of evidence – and, largely, of trust of the prosecutor ’s professionalism.⁶⁴

This means that, even if reliance on the public prosecutor ’s quasi-judicial determination is a strong enough basis for outweighing the presumption of innocence, that rationale cannot be applied to the first and other early remand hearings.

As the various arguments of principle fail, we come ever closer to

^{Antony Duff’s argument that, if we are serious about the presumption}

⁶² See eg s 10 of the Bail Act 2000 (NZ) and s 25 of the Criminal Justice and Public Order Act 1994 (UK) (as amended). However, even if such remands are thought justifiable in cases of serious violence, the presumption of innocence should operate to disallow the placement of a burden of proof on the defendant.

⁶³ Osman v United Kingdom (1999) 29 EHRR 245, and subsequent decisions discussed in Emmerson, Ashworth and Macdonald, above n 52, ch 19B. See also Van Colle v United Kingdom (2013) 56 EHRR 839.

⁶⁴ The English Code for Crown Prosecutors (2010), section 5, recognises this explicitly with a lesser ‘threshold’ test for early hearings involving people suspected of serious offences: see <www.cps.gov.uk>.

of innocence, then custodial remands before trial are absolutely unacceptable, chiefly because they treat defendants as objects or risk-bearers rather than as autonomous agents not proven to harbour harmful intentions.⁶⁵ Many governments would find this an unattractive conclusion; but that raises once more the question of the relationship between pre-trial detention and the presumption of innocence. The relationship is not straightforward, in view of the fact that remand proceedings do not involve the determination of guilt or the imposition of punishment. However, pre-trial deprivation of liberty involves imprisonment, often in poor conditions, ⁶⁶ and it could be argued that (despite the preventive purpose) pre-trial detention fulfils or comes close to fulfilling the definition of punishment. It inflicts material deprivations, including extreme restrictions on freedom of movement; low levels of comfort and amenity; idleness, with few opportunities for paid labour; relative isolation from family members, friends and the wider community; significant loss of autonomy in everyday life; substantial loss of privacy’ and exposure to the risk of personal harm.⁶⁷

Can the imposition of such deprivations on a person who ought to retain the presumption of innocence be justified? Are the right of personal liberty and the presumption of innocence sufficiently safeguarded if one of the three or four justifications for pre-trial detention is established by cogent evidence, and the court applies the principle of the least restrictive alternative (no lesser method of protection, such as bail with conditions, would be adequate)?

Answers to these questions are rendered more awkward by the presence of many guilty pleaders: in England and Wales around a half of all persons charged with offences indicate from an early stage their intention to plead guilty, and in this context it may be thought artificial to rely on the presumption of innocence as a reason against remanding them in custody. If it is right to assume that the early indication of a guilty plea amounts to a free admission in almost all cases (doubts might be harboured about some of those who later change their plea from not guilty to guilty), are guilty pleaders not accepting that there is good reason to treat them as guilty? On this view, if their offending behaviour is such as would normally lead to a custodial sentence, little is lost by remanding them in custody pending their court hearing – or, at least, that issues about the presumption of innocence are of less pressing

⁶⁵ R A Duff Trials and Punishments (Cambridge University Press, Cambridge,

^{1986) at 140, and now R A Duff “Pre-Trial Detention and the Presumption}

^{of Innocence” in A Ashworth, L Zedner and P Tomlin (eds) Prevention and}

^{the Limits of the Criminal Law (Oxford University Press, Oxford, 2013).}

^{66 In England and Wales remand prisoners often have to put up with the}

^{worst of prison conditions, an anomaly highlighted by Lord Windlesham}

^{in “Punishment and Prevention: The Inappropriate Prisoners” [1988]}

^{Criminal Law Review 140.}

^{67 This list is adapted from R Lippke “No Easy Way Out: Dangerous}

^{Offenders and Preventive Detention” (2008) 27 Law and Philosophy 383}

^{at 408.}

significance. In order to sustain this reasoning, courts should be legally required to take account of the likely sentence and not to remand a person in custody if they are unlikely to receive a custodial sentence, or more than a short sentence. Now that courts are accustomed to giving advance indication of sentence,⁶⁸ this kind of assessment should be perfectly possible. However, even if this reasoning were conceded, it can have no effect on the appropriate use of pre-trial detention for defendants whose offences would not merit imprisonment and, more importantly, for those who plead not guilty – indeed, it highlights the significance of their position, in relation to the presumption of innocence.⁶⁹

It is not helpful to assert, as did the US Supreme Court, that pre-trial detention is “essentially regulatory” and imposed in order to protect “the safety of the community”,⁷⁰ since that begs questions about the foundations of the state responsibility that is being asserted. True, the remand proceedings give the state the opportunity (through the courts) to take some such protective action, but are there strong enough reasons for thus overriding the right to personal liberty and the presumption of innocence? For many years the Irish courts maintained that the presumption of innocence prevented pre-trial detention motivated by a desire to reduce offending by defendants granted bail, and indeed that pre-trial detention treated the current charge as indicative of guilt.⁷¹ The strongest reasons identified above are the importance of ensuring that a defendant is called to account at a public trial (which would justify pre-trial detention if there were substantial grounds for believing that the defendant would abscond and this were the least restrictive means of securing the defendant’s attendance at trial), and the importance of securing the integrity of the administration of criminal justice (which would justify pre-trial detention if there were substantial grounds for believing that the defendant would interfere with witnesses and this were the least restrictive means of ensuring that such interference did not occur). A further reason would be the protection of a particular person from harm, which would justify pre-trial detention if there were substantial grounds for believing that the defendant would do harm to

68. In England, the regime for advance indications of sentence was laid down by the Court of Appeal in R v Goodyear [2004] EWCA Crim 1197; [2006] 1 Cr App R (S) 23.
69. See R L Lippke The Ethics of Plea Bargaining (Oxford University Press, Oxford, 2011), passim and, on innocent defendants, 82–84 and 212–216.

^{70 United States v Salerno [1987] USSC 94; (1987) 481 US 739; see also R v Hall [2002] 3 SCR}

^{309, where the Supreme Court of Canada held that it is not contrary to}

^{the Charter to deny bail if “necessary in order to maintain confidence}

^{in the administration of justice.” Cf the discussion of the nature of the}

^{power of pre-trial remand in the High Court of Australia in Chu Kheng}

^{Lim v Minister for Immigration, Local Government and Ethnic Affairs (1992)}

^{176 CLR 1 at 28, and the analysis by S McDonald “Involuntary Detention}

^{and the Separation of Judicial Power” [2007] FedLawRw 2; (2007) 35 Federal Law Review 25,}

^{esp at 72.}

^{71 U ni Raifeartaigh “Reconciling Bail Law with the Presumption of}

^{Innocence” (1997) 17 Oxford Journal of Legal Studies 1.}

that person unless detained. Apart from that, and for other exceptional categories, the general argument relating to future offending in general seems weak, both in its rationale and in its evidential basis. Liberty would be sacrificed on the basis of ungrounded predictions. And the argument that the decision of a public prosecutor to charge the defendant makes his subsequent conduct the proper concern of the government seems to be simply inconsistent with the presumption of innocence. Uncomfortable as it is, therefore, the conclusion is that there are not strong enough grounds for the pre-trial detention of defendants predicted as likely to commit offences, at least if they are not pleading guilty to offences for which imprisonment is to be expected.⁷²

IV When, if at all, can Indeterminate Custodial Sentences be

^Justified?

Criminal sentences that are proportionate to the seriousness of the offence can be justified on retributive grounds, and indeed many non-retributive rationales for punishment accept that there should be a limit on sentences in order to avoid disproportionality. Much of this terrain is contested, as is the meaning and measure of “proportionality” itself,⁷³ but our focus here is on a widely-recognised exception. Most legal systems have some kind of indeterminate sentence of detention for “dangerous” offenders, such as the sentence of preventive detention in New Zealand. In English law, for example, there are currently two main sentences of this kind (life imprisonment and extended sentences), but there are many people still in prison under a form of sentence (imprisonment for public protection (IPP)) which was abolished in 2012. Both life imprisonment and the former IPP are indeterminate sentences consisting of three stages: first, the minimum term or tariff period, set by the court in proportion to the seriousness of the offence(s); secondly, the public protection period, the length of which is determined by the Parole Board, which should not permit the offender ’s release until satisfied that there is no more than a minimal risk of further dangerous offences; and thirdly, the period of release on licence for the remainder of the offender ’s life, subject to recall to prison.⁷⁴ Both life imprisonment and IPP are severe sentences, depriving an individual of liberty for an indeterminate and usually very long period. This is not the place to describe the somewhat turbulent history of the IPP sentence;⁷⁵ suffice it to say that it was introduced in

2003 with a wide ambit that required the courts to impose IPP sentences on many offenders, some of whose offences were not particularly

⁷² See also Butler and Butler, above, n 16 at 743.

^{73 See, eg, A von Hirsch, A Ashworth and J Roberts (eds) Principled Sentencing}

^{(3rd ed, Hart, Oxford, 2009) ch 4.}

^{74 See N Padfield, D van Zyl Smit and F Dunkel (eds), Release from Prison:}

^{European Policies and Practices (Willan, Cullompton, 2010).}

^{75 For an excellent analysis, see J Jacobson and M Hough Unjust Deserts:}

^{Imprisonment for Public Protection (2010) <www.prisonreformtrust.org.}

^uk>.

serious; and that the mandatory provisions were removed in 2008, and the application of IPP sentences restricted to more serious offenders. In

2012 Parliament abolished the IPP sentence in favour of an automatic life sentence for the second serious qualifying offence, intended to cover fewer and more serious offenders.⁷⁶

The avowed purpose of indeterminate sentences such as life imprisonment and IPP is public safety. Once the life or IPP sentence has been imposed and the minimum term served, the task of public protection passes to the Parole Board, which is required not to authorise the release of such prisoners until they cease to pose a more than minimal threat of sexual or violent offending.⁷⁷ Thus the second and third stages of the life or IPP sentence are driven by considerations of public protection, which in this context means the prevention of serious violent or sexual offences. However, normative criticisms apart, there is a major problem with the internal logic of this rationale. This is because the (indeterminate) deprivation of liberty authorised by the sentences of life imprisonment and IPP is undoubtedly severe, but the empirical foundations for the predictions on which it is based are fragile and unconvincing. Research continues to show that predictions are more frequently wrong than right,⁷⁸ which means that the public is receiving much less protection than is supposed, and that over half of these indeterminate sentences turn out to have been unnecessary. The fallibility of predictions of this kind is so well established that even the current British coalition government has admitted that “the limitations of our ability to predict future serious offending ... [call] into question the whole basis on which many offenders are sentenced to IPPs.”⁷⁹ Taken at face value, as it surely should be, this observation largely undermines the case for preventive sentences of the kind being discussed here.

If the arguments are so badly flawed, can indeterminate preventive sentences be rescued by any deeper justifications? It is surely not controversial that one of the duties of the State is “to seek to reduce the incidence of the kinds of conduct that are properly criminalized, since it is a proper part of the state’s responsibility to seek to protect its citizens from suffering such wrongs.”⁸⁰ Although this duty contains at least two major loci of indeterminacy – how much of a reduction? which harms? – we can quickly reduce the significance of the second one by confining the present discussion to the prevention of serious violent or sexual harms.

⁷⁶ Legal Aid, Sentencing and Punishment of Offenders Act 2012 (England and Wales). However, over 5,000 offenders remain in prison subject to IPP sentences, which is why they are discussed here.

⁷⁷ For the original proposal, see Home Office Justice for All: A White Paper on the Criminal Justice System, CM 5563 (TSO, London, 2002) at [5.41].

^{78 See eg J L Skeen and J Monahan “Current Directions in Violence Risk}

^{Assessment”, (2011) 20 Current Directions in Psychological Science 38.}

^{79 Ministry of Justice Breaking the Cycle: Effective Punishment, Rehabilitation}

^{and Sentencing of Offenders, CM 7972 (TSO, London, 2010) at [186].}

^{80 R A Duff Answering for Crime: Responsibility and Liability in the Criminal}

^{Law (Hart Publishing, Oxford, 2007) at 87.}

But we are still faced with uncertainty about the degree of reduction of frequency that should be sought, or, differently put, the appropriate level of protection from these harms. Given the fallibility of predictions of such serious harms, some have turned to the “precautionary principle,” used chiefly in environmental law to respond to data gaps and uncertainties about risk in relation to possibly catastrophic consequences.⁸¹ On the precautionary principle, uncertainty about risk is not a reason for failing to take action, if the possible consequences of such failure might have significant and permanent effects. Limitations on our knowledge and predictive power should not stand in the way of taking precautions designed to avert possible catastrophic harms.

Is the precautionary principle appropriately deployed here? First, while the grounds for invoking the precautionary principle might appear to be satisfied by the low rate of successful predictions and the consequent uncertainties in relation to life imprisonment and the former IPP sentence, there is a major difference in respect of what can be justified by the principle. Sentences of preventive detention impose indeterminate deprivation of liberty on individuals. This, assuming the abolition of capital punishment, is the most severe permissible sentence and also the greatest permissible invasion of a person’s civil rights. There is nothing in the precautionary principle that is sufficiently powerful and persuasive to justify such a swingeing deprivation of liberty, such a deep invasion of an individual’s autonomy; to put the matter differently, no convincing argument has been advanced for depriving an individual of so much liberty and so many rights when the prospects of prediction are so poor.

Secondly, however, there are those who have argued that, once a person has committed one serious sexual or violent offence, that person has lost the right to be treated as free from harmful intentions when sentenced for a second such offence. Thus Floud and Young argued for a “just redistribution of risk” between people who have demonstrated their capacity to perpetrate serious wrongs, and their potential future victims. While other members of society have the benefit of the presumption of innocence and should be presumed harmless, those who have previously committed such a serious crime can properly be made subject to restrictions, so as to reduce the risks to potential victims.⁸²

^{Why might deprivation of liberty be easier to justify if the person has}

⁸¹ From a burgeoning literature, see E Fisher “Laws of Fear: Beyond the Precautionary Principle” (2006) 69 Modern Law Review 288; L Zedner “Fixing the Future? The Pre-Emptive Turn in Criminal Justice” in B McSherry, A Norrie and S Bronitt (eds) Regulating Deviance (Hart Publishing, Oxford,

2009); B Hebenton and T Seddon “From Dangerousness to Precaution: Managing Sexual and Violent Offenders in an Insecure and Uncertain Age” (2009) 49 British Journal of Criminology 343; P Ramsay “Imprisonment under the Precautionary Principle” in I Dennis and G R Sullivan (eds) Seeking Security: Pre-empting the Commission of Criminal Harm (Hart Publishing, Oxford, 2012).

⁸² J Floud and W Young Dangerousness and Criminal Justice (Heinemann, London, 1981).

already committed a serious crime? Antony Duff has argued that the presumption of harmlessness is not unconditional, and that it can be undermined by the commission of serious offences.⁸³ A similar approach is taken by Alex Walen: the State should presume people to be law- abiding, but a person who has been convicted of a very serious crime loses that status, and may therefore be subjected to preventive detention beyond the retributive sentence.⁸⁴ However, this way of viewing the issue says nothing about the poor prospects of prediction.

It is possible that some such forfeiture theory underlies the latest English protective sentence – an automatic life sentence for the second serious sexual or violent offence, the threshold being that the crimes must be ‘worth’ a 10-year sentence at least.⁸⁵ That precondition is likely significantly to limit the number of offenders eligible for the automatic life sentence, which is appropriate in view of the uncertainties (degree of gravity of predicted conduct; degree of probability of predicted conduct). Nonetheless, this new measure will authorise the deprivation of liberty beyond that justifiable on proportionality grounds, in the face of the uncertainties outlined above.

Even if a “forfeiture of rights” argument along the above lines can be sustained, the absence of reliable predictive evidence makes it problematic to determine how much more deprivation of liberty, over and above what is proportionate to the offence committed, can be justified.⁸⁶

Moreover, if and insofar as a legal system justifies detention beyond the proportionate sentence on preventive grounds, such preventive detention should not be served in the same conditions as (punitive) imprisonment. There are strong arguments in favour of requiring any preventive sentence to be served in a non-prison institution, with treatment programmes available.⁸⁷

V Negotiating Liberty

Both the New Zealand Bill of Rights Act and the European Convention on Human Rights recognize the right to liberty of the person, and stipulate that it is not to be taken away arbitrarily. In this lecture I have

⁸³ R A Duff “Dangerousness and Citizenship” in A Ashworth and M Wasik

^{(eds) Fundamentals of Sentencing Theory (Oxford University Press, Oxford,}

^{1998) 152.}

^{84 A Walen “A Punitive Precondition for Preventive Detention: Lost Status}

^{as a Foundation for Lost Immunity” (2011) 48 University of San Diego Law}

^{Review 1229.}

^{85 See n 76 above, and text.}

^{86 See the British government’s recognition of this, above n 79; and further,}

^{A von Hirsch and A Ashworth Proportionate Sentencing: Exploring the}

^{Principles (Oxford University Press, Oxford, 2005) at 51–61.}

^{87 This is effectively the result of M v Germany (2010) 51 EHRR 976, where}

^{the relevant German provision was found to violate the European}

^{Convention. Similar requirements have been put forward in judgments}

^{such as James, Wells and Lee v United Kingdom (2013) 56 EHRR 399 and}

^{Vinter v United Kingdom judgment of 9 July 2013.}

argued that this fundamental right to liberty has a significant penumbra of uncertainty, to the extent that its practical impact may be reduced significantly by preventive considerations. I recognize that there must be exceptions to the right to liberty, in order to allow the law enforcement authorities to take measures to ensure public safety, to prevent crimes and to investigate crimes that have already happened. That forms part of what I have suggested is a democratic compromise, in which citizens yield a certain amount of liberty in order to receive a measure of state protection from harms. Where the law enforcement authorities exercise one of those powers, the affected persons should benefit from special rights and safeguards (eg when arrested or detained).

Any notion of a democratic compromise begins to look less convincing, however, to the extent that the ambit of one’s liberty becomes uncertain, as John Stuart Mill recognised many years ago.⁸⁸ Thus we noted the uncertainties surrounding arrest and detention by the police – the European Court’s emphasis on (background) coercion, and the existence of reasonable grounds for believing that one’s freedom of movement no longer exists. Much as the law enforcement authorities may wish to portray a conversation with a citizen as “voluntary”, the presence of a coercive umbrella may give grounds for believing that there is an arbitrary arrest or detention. We then considered the use of coercive tactics in public order policing, notably the device of “kettling.” Although the common law justification for this tactic is said to lie in the power to prevent a breach of the peace, we noted the arrival of justifications claiming that the right to life – the most fundamental right of all – was engaged and therefore in conflict with the right to personal liberty in such cases. The Grand Chamber ’s judgment in the Austin case not only says little about this approach, but is also disappointing in its denial that being “kettled” amounts to a deprivation of liberty. This propagates confusion about the relative significance of the purpose and the degree of liberty-constraining situations. At least it can be agreed that the courts must insist on the principle of the least restrictive alternative, rightly promoted in the G20 judgments.⁸⁹

Turning to pre-trial detention, I suggested that both the right to liberty and the presumption of innocence are called into question by the assumption – and I think it is an assumption – that the state bears responsibility for a defendant’s conduct once that person has been charged with an offence. On what is this responsibility founded? If there is evidence of a threat to a particular person, or a risk of absconding or of interfering with witnesses, that is one thing. But if the prosecutor ’s

⁸⁸ Mill wrote that the preventive argument is “far more liable to be abused, to the prejudice of liberty, than the punitory function; for there is hardly any part of the legitimate freedom of action of a human being that would not admit of being represented, and fairly too, as increasing the facilities for some form or other of delinquency”: J S Mill On Liberty (J W Parker and Son, London, 1859) at 106.

^{89 See above, n 44–47 and accompanying text.}

decision to charge – before any evidence has been presented – is put forward as the basis for depriving someone of their liberty, should that really be enough to overcome the presumption of innocence? And if defendants who indicate an intention to plead guilty are brought into the calculation, should the court not be required to determine that a significant custodial sentence would be likely, before pre-trial detention is authorised? Part of the problem here is the relatively low success rate in predicting (further) offences, and that becomes an issue also in relation to preventive detention in the form of indeterminate custodial sentences. We considered the suggestion that a person who has already been convicted of a serious violent offence should no longer have the benefit of a cousin of the presumption of innocence, the presumption of harmlessness. Even if that is persuasive, we are left with the poor rates of prediction compared with the substantial deprivation of liberty. Any such preventive detention would, like pre-trial detention, fall outside the definition of punishment, and it is therefore right that it should be differentiated in practice by means of different institutions and different facilities. What remains of the right to liberty of the person is surely worth that.