Persistence, prevention and prediction

Terms such as ‘professional criminals’ and ‘real menaces’ have been usedwithout much effort at precision, and when the law did eventually specify violentand sexual offenders, many of tho

CHAPTER 6

Persistence, prevention and prediction

In the course of this chapter there will be detailed discussion of a group of sentencingissues which lead to considerable practical and theoretical difficulties After a briefhistorical introduction, part2 explores three possible approaches to sentencingpersistent offenders, and part3considers the relevant provisions of the CriminalJustice Act 2003 Parts4and5examine two specific problems, those of ‘professional’criminals and of petty persistent offenders In Part 6 a new approach to prevention

is examined – the use of anti-social behaviour orders Part7of the chapter turns tothe question of selective incapacitation as a strategy for preventing crime, referring

to the minimum sentences in English law In part8the ‘dangerousness’ provisions

of the 2003 Act are examined, and some concluding thoughts are found in part9.Throughout these topics there are linking themes concerned with the promotion

of security and the assessment of risk of future criminal behaviour The invocation

of such rationales amounts to a departure from the proportionality principle, andclose attention will be paid to the justifications for this

6.1 Historical introduction

The history of English measures aimed specifically at persistent offenders seems

to be widely acknowledged to be a history of failure The judges have had cient discretion, for the last hundred years at least, to allow them to pass fairlylong sentences on persistent serious criminals without invoking any special pow-ers But penal reformers and governments have invariably felt that no major set ofreforms would be complete without making further special provision for persistentoffenders The Gladstone Committee in 1895 argued in favour of a special measureagainst persistent thieves and robbers, who would otherwise serve a succession offairly short sentences and therefore return frequently to prey on the community.The Committee’s proposals led, after much debate,1to the Prevention of CrimeAct 1908 This empowered a court to impose, upon an offender with three pre-vious felony convictions since the age of 16, a sentence of preventive detention ofbetween five and ten years, in addition to the normal sentence for the crime (a

suffi-1 Radzinowicz and Hood ( 1986 ), pp 265–78.

182

so-called ‘double track’ system) The practical focus of the Act was soon revisedwhen Churchill became Home Secretary He took the view that the Act, as it wasbeing administered, concentrated unduly on mere repetition in lawbreaking, and

he exposed the minor nature of some of the offences which had led to the sition of preventive detention He issued a new circular which declared that ‘merepilfering, unaccompanied by any serious aggravation, can never justify’ preventivedetention, and propounded the general test of whether the nature of the crime was

impo-‘such as to indicate that the offender is not merely a nuisance but a serious danger

to society’.2The aim of preventive detention thus became that of ‘protecting societyfrom the worst class of professional criminal’ In fact, the courts often found theirordinary sentencing powers sufficient in such cases, and so the use of preventivedetention declined

A new form of preventive detention was introduced in the Criminal Justice Act

1948 for persistent offenders aged 30 or over, being a sentence of 5 to 14 years instead

of (not in addition to) the normal sentence The Dove-Wilson Committee in 1932had proposed this as suitable chiefly for ‘professional criminals who deliberatelymake a living by preying on the public’,3but when the legislation was introduced in

1948 the government envisaged that it would also cover ‘the relatively trivial sistent] offender’.4Judges soon found themselves passing sentences of preventivedetention on offenders whose records, while showing persistence, were not serious

[per-In the late 1950s the judges increasingly set their faces against this, and in 1962 theLord Chief Justice went so far as to issue a Practice Direction to restrict the use ofpreventive detention.5Following a gloomy report from the Advisory Council onthe Treatment of Offenders (1963) and two other studies which demonstrated theminor nature of many of the offences committed by those subjected to preventivedetention,6the sentence virtually fell into disuse

The next measure to be introduced was the extended sentence: the CriminalJustice Act 1967 empowered a court to extend a sentence beyond the normal length

or (in limited circumstances) beyond the statutory maximum where it apprehendedthe need, in view of the offender’s record, to protect the public The White Paper of

1965 had proposed the extended sentence for those offenders who constitute ‘a realmenace to society’,7but a parliamentary amendment which would have requiredthe court to have regard to the gravity of the current offence was not accepted by thegovernment Once again the courts soon found that those falling within the ambit

of the sentence could hardly be described as real menaces, and at no stage did theextended sentence play a significant part in sentencing practice

In some respects the Criminal Justice Act 1991 was a slight improvement, since

s 2(2)(b) permitted courts to pass a longer than proportionate sentence for a violent

or sexual offence if it was of opinion ‘that only such a sentence would be adequate

2 Radzinowicz and Hood ( 1986 ), p 285 3 Dove-Wilson ( 1932 ), para 42.

4 Hammond and Chayen ( 1963 ), p 11 5 See [1962] 1 All ER 671.

6 Hammond and Chayen ( 1963 ), West ( 1963 ) 7 Home Office ( 1965 ).

to protect the public from serious harm from him’ The Act did attempt to define

‘violent’, ‘sexual’ and ‘serious harm’, but the provision was vague on crucial issuesand unacceptably wide in its scope as interpreted by the courts.8 The power toimpose longer than proportionate sentences appears to have been used relativelyrarely in recent years, perhaps (again) because courts tend to give long sentencesfor serious offences anyway Thus a brief historical survey reveals two recurrentdifficulties First, legislation on persistent offenders has usually been framed inbroad terms, often without clear and precise guidance about the types of offender

to be included and excluded Second, and more fundamentally, there has been littleagreement about the group or groups of offenders who should be the target of specialsentences Terms such as ‘professional criminals’ and ‘real menaces’ have been usedwithout much effort at precision, and when the law did eventually specify violentand sexual offenders, many of those included were at the lowest end of the scale.6.2 Three approaches to punishing persistence

The differing views which have been expressed on the sentencing of persistentoffenders do not always fall neatly into categories, but three paradigms are (i) flat-rate sentencing, (ii) the cumulative principle, and (iii) progressive loss of mitigation.The aim here is to offer a description of each approach, to consider its rationale,and to weigh the advantages and disadvantages.9

6.2.1 Flat-rate sentencing

According to this approach, the sentence should be governed by the crime and not

at all by the offender’s prior record This view has been advanced by a small group

of desert theorists, most notably George Fletcher10 and Richard Singer.11 Theirargument, in brief, is that an offender’s desert should be measured by reference tothe crime committed, in terms of its harmfulness and the offender’s culpability inrelation to it Any previous offences cannot have a bearing on this Indeed, not onlyare they irrelevant to the calculation, but to take them into account would be topunish the offender twice over – if sentence has already been passed for the previousoffences, it is unjust to increase the sentence for a subsequent crime on account of

a previously punished offence Fletcher suggests that desert theorists who do takeaccount of previous offences are indulging in a covert preventionist strategy Sincethe increased sentence cannot be justified on desert grounds, says Fletcher, suchwriters are really trying to achieve a modest amount of individual prevention orincapacitation in such cases

There are few practical examples of flat-rate sentencing schemes Illegal parking

of cars is one: the penalty does not increase according to the number of previousoffences, and one could commit the offence every day without ever receiving more

8 See the 3rd edn of this work, pp 183–9 9 For fuller discussion see Roberts ( 1997 ).

10 Fletcher ( 1978 ), pp 460–6 11 Singer ( 1979 ), ch 5.

than the fixed penalty Many other minor offences have fixed penalties or such lowmaximum fines that they may be viewed as flat-rate offences If a case goes to court,that creates an opportunity for the penalty to be mitigated for a poor or first-timeoffender – an opportunity ruled out by flat-rate sentencing Indeed, since there can

be no concession for the impecunious or first offender, the practical question arises

of the level at which flat-rate penalties might be set if they are to be defended asproportionate

6.2.2 The cumulative principle

Since at least the mid-nineteenth century there has been support for the lative principle of sentencing persistent offenders The basic idea is that, for eachnew offence, the sentence should be more severe than for the previous offence

cumu-In this way sentences should be cumulative, with a view to deterring the ual offender from repeating the crimes Perhaps the best-known exponent of thecumulative principle was the Gloucestershire magistrate Barwick Lloyd Baker In

individ-1863 he proposed that for a first felony conviction the punishment should be oneweek or ten days’ prison on bread and water; for the second conviction 12 months’imprisonment; for the third, seven years’ penal servitude; and for a fourth, penalservitude for life or for some very long period which would allow surveillance onticket-of-leave for the greater part of the criminal’s life He saw this as achievingprotection through individual deterrence, and had no doubts about its fairness: ‘ifyou tell a man clearly what will be the punishment of a crime before he commits it,there can be no injustice in inflicting it’.12

That harsh approach made no allowances for the fact that some offences wereminor and some stemmed from human weakness or poverty rather than ‘wicked-ness’ It met with considerable opposition, notably from Francis Hopwood, Recorder

of Liverpool towards the end of the nineteenth century, who strenuously denouncedheavy penalties for petty recidivists The Lord Chief Justice of the time, LordColeridge, appeared to have had greater sympathy with Hopwood’s approach, since

he maintained that he would inflict punishment only ‘for the particular offence forwhich the prisoner is being tried before me’ But even Lord Coleridge admitted thatsome of his colleagues had ‘different guiding thoughts’.13

Although Baker’s rationale was deterrence, incapacitation might also be invoked

in support of the cumulative approach, especially in view of the contemporaryemphasis on security and risk If so, how might one identify the offenders againstwhom the principle should be applied? Or is it contended that society is justified inprotecting itself against all persistent offenders? Baker had argued that cumulativesentencing of habitual misdemeanants would reduce the incidence of petty offences

by some 60,000 a year.14There are Home Office statistics showing that offenders withfive or more previous convictions are 87 per cent likely to be convicted of another

12 See Radzinowicz and Hood ( 1986 ), pp 237–8 and references.

13 Radzinowicz and Hood ( 1979 ), pp 1311–12 14 Radzinowiz and Hood (1980), p 1330.

offence within six years.15However, most of these offences are towards the lower end

of the scale of criminality: the high rates of recidivism are for lesser crimes, and thecumulative principle therefore tends to heap punishment on minor and relativelynon-threatening offenders The second and related question is whether, even ifthere was evidence that cumulative sentencing would ‘work’, it would satisfy therequirements of fairness Would it be acceptable if the sentencing system prescribedpenalties of ascending severity for recidivists, especially if their offences were minor?Would it be any more acceptable if the extra detention was under non-punitiveconditions, perhaps similar to quarantine? These fairness issues are taken up in part7

of this chapter, where proposals for the selective incapacitation of certain types ofoffender are reviewed

The more common rationale for cumulative sentencing is individual prevention.This was Baker’s main argument: cumulative penalties would deter the offender or,

if they did not, he would in effect ‘with his eyes open deliberately sentence himself ’.16

Several questions are raised by this claim Are all, or even most, persistent offendersthe rationally motivated wicked offenders that it assumes? The historical evidence

of measures against persistent offenders, reviewed briefly in part1of this chapter,suggests that many of them are not.17 Among them are people who are sociallydisadvantaged, others who are in personal turmoil, and others who are mentallydisturbed Even if such offenders are regarded as a threat to the public, the cause oftheir offending indicates that more constructive measures should be taken, and theconcept of fairness underlying the proportionality principle should place a limit tothe power that may be taken over them For dangerous offenders it may be arguedthat people who offend repeatedly forfeit any right to be regarded as full rights-bearing members of society, or that any rights which such people have ought to beset against the rights of their potential victims, so that the rights of a person whohas been shown to reoffend repeatedly may justifiably be overridden in order topreserve the rights of others These and other arguments are considered in part8

of this chapter But for the socially disadvantaged or mentally disturbed, this is not

an appropriate standpoint

Would the cumulative strategy be effective in preventing crime? This dependsnot only on such factors as knowledge of the penalties among offenders and theabsence of countervailing considerations (e.g low detection rate, absence of propersocial provision for people in need), but also on the effectiveness of the penalty.English law has long had one form of cumulative sentencing – the penalty pointssystem for road traffic offenders When a court sentences an offender for certaintraffic offences, it may (or must) impose a number of penalty points, and when anoffender accumulates 12 points an immediate disqualification from driving follows.The justifications for having this system for motoring offences and not for other

15 Philpotts and Lancucki ( 1979 ), p 16; see also Lloyd, Mair and Hough ( 1994 ).

16 Radzinowicz and Hood ( 1986 ), p 238.

17 Radzinowicz and Hood ( 1986 ), chs 8–12, on these debates in the nineteenth century.

crimes have yet to be debated widely; but when it comes to a preventive systembased on sentences of imprisonment, the objection was pointed out as long ago as

1932by the Dove-Wilson Committee:

the inference is that present methods not only fail to check the criminal propensities ofsuch people, but may actually cause progressive deterioration by habituating offenders

to prison conditions which weaken rather than strengthen their characters.18

Thus the repeated use of prison sentences may be counter-productive, making theseoffenders less able to live law-abiding lives and more likely to reoffend on release

If the cumulative principle is based on individual deterrence, and if the point ofdeterrence is to protect the public, heavy reliance on imprisonment for this purposemay not only go against the principle of restraint (see Chapter3.4.2) but also be to

a significant extent self-defeating

On the basis of a review of the available evidence some ten years ago on thetypical characteristics of criminal careers, David Farrington argued that ‘since ahigh proportion of offenders desist after the first or second offence, significantcriminal justice interventions might be delayed until the third offence Diversionarymeasures might be appropriate after the first or second offence.’19 This drives afurther wedge between cumulative sentencing and prevention If prevention is to

be the chief concern, it does not follow that cumulative sentencing is the mosteffective way of achieving this, particularly after two or three convictions, andparticularly if incarceration is involved Like many penal policies, it may have asuperficial attractiveness to politicians and the media because it appears ‘tough’,but it relies on crude assumptions about the causes of offending and on a failure tograsp the criminogenic effects of the penal system itself.20

6.2.3 Progressive loss of mitigation

This approach to the sentencing of persistent offenders differs from flat-rate ing in making some allowance for previous record, and differs from the cumulativeprinciple in placing limits on the influence of previous record and in deferring to anoverall concept of proportionality The principle of progressive loss of mitigationreally consists of two parts: one is that a first offender should receive a reduction

sentenc-of sentence, and the other is that with second and subsequent sentenc-offences an sentenc-offendershould progressively lose that mitigation How soon all the mitigation is lost is aquestion for discussion later, but clearly the principle assumes a limit beyond whichthe sentence cannot go, no matter how many previous convictions the offender has.The gravity of the current offence(s) is taken to set a ‘ceiling’ for the sentence: a badprevious record should mean that the offender loses this source of mitigation, butthe record should not be treated as an aggravating factor As Thomas put it, a bad

18 Dove-Wilson ( 1932 ), para 3 19 Farrington ( 1997 ), pp 564–5.

20 On the indices of risk of offending and how to respond to them, see Farrington ( 2002 ).

record ‘will not justify the imposition of a term of imprisonment in excess of thepermissible ceiling for the facts of the immediate offence’.21

What is the theory underlying progressive loss of mitigation? It is an approachcharacteristically adopted by desert theorists, who view proportionality to the seri-ousness of the offence as the chief determinant of sentence We saw earlier that somedesert theorists adopt flat-rate sentencing for recidivists Why would a desert theo-rist wish to dilute an offence-based system of sentencing (harm plus culpability)

by incorporating an element relating to the offender’s past history? The argument,restated and refined by Andrew von Hirsch,22is based on the idea of a lapse Ordi-nary people do have occasional aberrations Human weakness is not so unusual Thesentencing system should recognize not only this, but also the capacity of people torespond to censure, and to ensure that their future conduct conforms to the law This

is embodied in the idea of giving someone a ‘second chance’.23So the justificationfor the discount for first offenders rests partly on recognition of human fallibility,and partly on respect for people’s ability to respond to the censure expressed in thesentence The justification for the gradual losing of that mitigation on second andsubsequent convictions is that the ‘second chance’ has been given and not taken: theoffender has forfeited the tolerance, and its associated sentence discount, becausethe subsequent criminal choices show insufficient response to the public censure Inprinciple, therefore, the second offence deserves greater censure than the first (unlessthere is good reason to indicate otherwise), and the third offence may be censuredfully But the seriousness of the offence must remain the primary determinant ofsentence, and therefore sentences imposed on repeat offenders should not cumulate

so as to lead to custodial terms greater than the current offence could justify.One possible counter-argument is that the notion of lapse appears to take noaccount of the possibility that a first offender might have planned an offence metic-ulously and might have been fully aware of the gravity of the wrongdoing However,the ‘second chance’ theory turns on the ability to respond to censure and pun-ishment, not on mere awareness of wrongdoing A second counter-argument isthat the justifications offered seem to assume that all offending is based on ration-

al choice, and to ignore the findings of criminological research Thus rational choice(of a kind) may be evinced by those who adopt a particular lifestyle, such as careerburglars.24 But some recidivism is largely a concomitant of going to particularplaces and associating with particular people, as with people who frequently becomeinvolved in violence associated with drinking in public houses.25Some may stem

21 Thomas ( 1979 ), p 41.

22 Von Hirsch and Ashworth ( 1998 ), ch 4.7; von Hirsch and Ashworth ( 2005 ).

23 Cf Bagaric ( 2001 ), ch 10.3.1, for the counter-argument that the moral notion of lapse is propriate for matters so serious as criminal convictions However, it is questionable whether every criminal offence is sufficiently serious to remove the moral force of the argument from lapse Bagaric also argues that the idea of a discount for first offenders is a subterfuge, and that in reality

inap-we are discussing the claim of previous convictions to operate as an aggravating factor For contrary arguments, see ch 5.1 above.

24 E.g Maguire ( 1982 ), Bennett and Wright ( 1984 ) 25 Walmsley ( 1986 ), pp 17–18.

from contacts made within penal institutions, where information is exchanged andalliances formed.26A large amount of recidivism may be associated with drug use.27

And, more generally, some is part of a cycle of social deprivation and/or personalturmoil, which may or may not be deepened by the experience of imprisonment,

as with the so-called petty persistent offenders.28Studies of desistance from crime,which focus on the circumstances in which offenders typically give up offending,have long indicated the relevance of stable relationships, a child, a job and otherprosaic factors in a person’s life.29In this context it is important to note two of therecommendations of the Council of Europe on sentencing:

D1 Previous convictions should not, at any stage in the criminal justice system, be usedmechanically as a factor working against the defendant

D2 Although it may be justifiable to take account of the offender’s previous criminalrecord within the declared rationales for sentencing, the sentence should be kept inproportion to the seriousness of the current offence(s).30

Proposition D1 emphasizes the importance of considering the reasons for ing in each case This does not present problems for desert theorists, for, as we saw

reoffend-in Chapter4.5, they can accept grounds for mitigation based on diminished ity, social deprivation and so forth Proposition D2 recognizes that, even in thosecountries where prevention is the primary rationale of sentencing, there should be aproportionality constraint in the sentencing of persistent offenders The great merit

capac-of the ‘second chance’ idea is that a clear principle capac-of fairness is accorded a tral place.31Thus, progressive loss of mitigation assumes that a second and a thirdoffence deserve greater censure, but it accords with the Council of Europe in leav-ing room for other responses if other explanations for reoffending seem persuasiveand in insisting on a firm proportionality constraint This last point distinguishes

cen-it clearly from the cumulative approach, which may result in long sentences forpersistent but non-serious offenders

At common law the Court of Appeal frequently restated the theory of sive loss of mitigation and the proportionality constraint on sentencing persistent

progres-offenders The leading case was probably Queen (1981),32where the offender hadcountersigned and attempted to cash a cheque for £50 belonging to someone else

He was sentenced to 18 months’ imprisonment The Court of Appeal, accepting that

he had probably been sentenced on his ‘appalling’ record of thefts and deceptions,reduced the sentence to allow his immediate release Kenneth Jones J held that it iswrong in principle to sentence an offender on his record:

The proper way to look at the matter is to decide on a sentence which is appropriate forthe offence for which the prisoner is before the court Then in deciding whether thatsentence should be imposed or whether the court can properly extend some leniency

to the offender, the court must have regard to those matters which tell in his favour;and equally to those matters which tell against him, in particular his record of previousconvictions.33

The gravity of the offence should set the ceiling, and even an appalling prior recordshould not take the sentence above it However, while the rhetoric of the courtsembraced the principle, the reality of sentencing recidivists was often different.Moreover, the courts often seemed unaware of the difference, and managed to say

one thing and do another in the same case In Bailey (1988)34the offender stoodconvicted of two offences – one was theft of several ladies’ nightdresses, which

he had seized from a shop and taken to his solicitor’s office nearby; the otherwas burglary of a hospital, in the form of taking four packets of frozen cod filletsfrom a hospital freezer The trial judge imposed two years’ imprisonment for thetheft, and 18 months consecutive for the burglary, totalling three-and-a-half years.The offender’s record was described by the Court of Appeal as ‘truly appalling’: itstretched back over 25 years, though most were ‘comparatively petty thefts’ The trialjudge evidently imposed the sentence in order to incapacitate Bailey for a lengthytime (a version of the cumulative principle), but the Court of Appeal held that thiswas wrong in principle Stocker LJ went on:

It is of course manifest that a convicted criminal’s past record forms part of the matrixupon which he falls to be sentenced Clearly no court would be likely to impose a sen-tence of imprisonment for a first offender of the same length that might be appropriatefor a person with a substantial criminal record To that extent the past record is a rele-vant factor to be taken into account On the other hand, as has often been said by thiscourt the sentence imposed must be related to the gravity of the offences in relation

to which it is imposed Whilst fully understanding the motive which impelled thelearned judge to impose a total sentence of three and a half years, we feel bound to saythat those sentences bore so little relationship to the gravity of the offences that evenhaving regard to the appalling background of this appellant, they cannot possibly bejustified.35

The court went on to reduce the sentences to 15 months for the theft, and threemonths consecutive for the burglary which yielded the frozen cod fillets

The statement of principle in this case is fairly clear Although it is often shortened

to ‘it is wrong to sentence on record’, the court shows appreciation of the relevance ofthe prior record, within limits, to the sentence But what are those limits? They weresupposed to be set by reference to the gravity of the offence(s), and to constitute theceiling beyond which the sentence should not go But the scarcity of clear guidelinesfor sentencing meant that ceilings were often somewhat plastic By what benchmark

is 15 months’ imprisonment a proper ceiling for a rather feeble theft of nightdresses,

33 Ibid., at p 255 34 (1988) 10 Cr App R (S) 231 35 Ibid., at p 233.

which resulted in the recovery of the stolen property fairly soon after the event? The

Court of Appeal’s decision in Bailey shows that the rhetoric of the courts has often

been different from the reality of their sentencing practice, and that the principle

of progressive loss of mitigation cannot operate without reasonably firm sentenceranges

The principle of progressive mitigation was not applied at common law to allcrimes, even in theory A different approach, more akin to flat-rate sentencing,

applied to grave crimes In his pioneering judgment in Turner (1975),36on ing levels for armed robbery and for grave crimes in general, Lawton LJ stated that

sentenc-‘the fact that a man has not much of a criminal record, if any at all, is not a powerfulfactor to be taken into consideration when the court is dealing with cases of thisgravity’ Similarly, the guideline judgment on rape states that ‘the defendant’s goodcharacter, although it should not be ignored, does not justify a substantial reduc-tion of what would otherwise be the appropriate sentence’.37If this restriction is to

be explained, it must be along the lines that little concession to human weaknessshould be made where there is egregious wrongdoing The usual ‘concession tohuman frailty’ implies that the offence can be seen as an unfortunate lapse, whereasthere is less room for compassion for those who succumb to the temptation to com-mit a grave crime Presumably this approach implies a kind of sliding scale, with thegeneral ‘concession to human frailty’ approach to first offenders gradually givingway to a harder line Rape has a starting point of at least five years’ imprisonment,and so perhaps it is at that level that the concession tapers off

6.3 Previous convictions and the Criminal Justice Act 2003

Section 143(2) of the 2003 Act introduces a new provision on the sentencing ofrepeat offenders As such, it takes its place in a panoply of measures introduced

by the government to tackle persistent offending The discussion here begins byanalysing official policy in respect of persistent offenders; it then examines the newlegislative provision in some detail; and finally it considers what impact the new law

is likely to have on sentencing practice, on the prisons, and on public protection.6.3.1 Policy on persistent offenders38

Sentencing policy for persistent offenders should be seen in the broader context ofthe criminal justice system In recent years the government has required local Crim-inal Justice Boards to develop schemes for targeting persistent offenders in terms

of investigation and detection, aiming at ‘prolific offenders who are responsible for

a disproportionate amount of crime’.39The Persistent Offender Scheme defined a

‘core persistent offender’ as someone aged 18 or over who has been convicted of six

36 (1975) 61 Cr App R 67 at p 91, discussed in ch 4.4.9 above.

37 Millberry [2003] 2 Cr App R (S) 142 at pp 152–3, adopting SAP, Advice to the Court of Appeal: Rape (2002), para 46.

38 For detailed discussion see Wasik ( 1987 ), Roberts ( 1997 ), and von Hirsch and Roberts ( 2004 ).

39 Home Office ( 2002 ), p 1.

or more recordable offences in the last 12 months A recent inspection report onthe scheme found that the most common offence by far was theft from shops, at

36 per cent of all those falling within the definition Only 5 per cent of qualifyingoffences were burglary in a dwelling.40This demonstrates an abject failure to under-stand the lessons of history in respect of policy on persistent offenders, as described

in part6.1above, and it is hardly surprising that the first recommendation made

in the Joint Inspection Report is that the definition be narrowed so as ‘to identify

a more limited number of priority offenders’.41The report therefore recognizes theconfusion between mere repeat offending and the idea of targeting priority offend-ers, but it stops short of indicating exactly who the priority offenders should be Itgoes on to suggest that these should be identified locally, but the lack of specificitysuggests that policy on persistent offending is still wallowing in a vitiating uncer-tainty The report states that the needs of the offenders currently included in thescheme are similar to those in the general offending population – notably, ‘problemswith thinking skills, drug misuse, employment training and education, accommo-dation, lifestyle, attitudes and finance’.42The role of alcohol and substance misuseshould be given even greater emphasis:

The 100,000 most persistent offenders share a common profile Half are under 21 andnearly three-quarters started offending between 13 and 15 Nearly two-thirds are harddrug users More than a third were in care as children Half have no qualifications atall and nearly half have been excluded from school Three-quarters have no work andlittle or no legal income

This important quotation comes not from a criminological textbook but from

a major Home Office framework document.43However, it has been ignored toofrequently by the Home Office since it was written The quotation also points up athird difficulty with the prolific offender policy – the reference to 100,000 persistentoffenders who commit about 50 per cent of serious crimes in any one year TheHalliday report was pessimistic about the crime-preventive effects of targeting thisgroup,44and the Carter report referred to evidence that each year some 40 per cent

of persistent offenders will stop offending without further official intervention andthat many of these are replaced by new persistent offenders Carter concluded that

‘historically, incapacitation is only associated with small falls in crime’,45and thusshared Halliday’s view that subjecting this group to longer sentences would nothave the large crime-preventive effects sometimes assumed Criminological researchshows that many criminal careers are short-lived, and mostly among males aged15–25, and that there is a variety of personal and social-structural explanations ofwhy offenders desist after a few years.46The constant renewal of the stock of repeat

40 Joint Inspection Report ( 2004 ), para 6.13.

41 Joint Inspection Report ( 2004 ), para 7.8.

42 Joint Inspection Report ( 2004 ), para 6.14 43 Home Office ( 2001 ), Appendix B, para B.7.

44 Halliday ( 2001 ), Appendix 6; cf Appendix 3, on criminal careers.

45 Carter ( 2003 ), pp 15–16 46 See Bottoms, Shapland et al ( 2004), p 370 and passim.

offenders is a well-known explanation of why incapacitative sentencing policies areless effective than some expect.

What, then, should be the policy in dealing with persistent offenders, and why?The Halliday review concluded that the proper approach to sentencing persistentoffenders was unclear and therefore probably inconsistent – a diagnosis that couldhardly be disputed, given the gap between the rhetoric of the common law and thepractice of the courts (set out in part6.2.3above) and given the extraordinary factthat the legislative provision on sentencing persistent offenders introduced by theCriminal Justice Act 1993 was never formally considered by the Court of Appealduring its 12 years in force How, then, should the law be ‘clarified’?

Clarification needs to be based on a clear presumption that sentence severity shouldincrease as a consequence of sufficiently recent and relevant previous convictions Thejustification for this modified principle is twofold A continuing course of criminalconduct in the face of repeated attempts by the State to correct it, calls for increas-ing denunciation and retribution, notwithstanding that earlier crimes have alreadybeen punished In addition, persistent criminality justifies the more intensive efforts

to reform and rehabilitate which become possible within a more intrusive and tive sentence As it happens, because previous convictions are a strong indicator ofrisks of reoffending, this presumption would also, coincidentally, take such risks intoaccount For all these reasons, the new presumption would serve to target resources

puni-on the offenders who commit a disproportipuni-onate amount of crime and are most likely

to reoffend The new presumption must be governed by the proportionality principle,

to avoid excessively severe and therefore unjust punishments To do this, clear lines demonstrating the ‘gearing’ between offence seriousness, seriousness of record,and bands of acceptable sentences, will be needed, building on the guidelines alreadyestablished or under development.47

guide-Four points may be made about this key passage First, it puts forward two reasonsfor modifying the proportionality principle, and with it the principle of progressiveloss of mitigation To argue that repeated convictions necessarily amount to somekind of defiance of the state and thus supply a justification for increased severity isdoubtful for a number of reasons Notably, it is doubtful because it assumes thatthe motivation for and causes of reoffending can properly be seen as defiance ofthe state and its efforts: the quotation from the Home Office set out above48showswhat kinds of offender are to be found in this group, and it is therefore wrong toassume that they are all rational calculators who are deliberately thumbing theirnoses at the state And it is doubtful, in any event, whether it is justifiable to treat anydefiance of the state as an aggravating factor: to treat ‘insubordination’ as a seriousevil in itself, aside from any harm resulting from or risked by the conduct, is a form

of authoritarianism inconsistent with proper democratic principles.49Halliday didnot argue that his recidivist premium was justified by reference to deterrence or

47 Halliday ( 2001 ), para 2.7 48 See text at n 43 above.

49 See von Hirsch and Roberts ( 2004 ), p 649.

to incapacitation His primary rationales appear to be denunciation and a version

of proportionality that treats an offender as deserving more for each subsequentoffence, both of which are connected with his flawed idea of defiance

Second, when the above passage does recognize that some persistent ers need rehabilitation, it insists that this should be carried out within ‘a moreintrusive and punitive sentence’ Granted that such a significant proportion of per-sistent offending is drug-related,50does this mean that drug treatment for persistentoffenders should generally be carried out in custody? Is that the right policy, in terms

offend-of either the seriousness offend-of the individual offend-offences or the effectiveness offend-of the ment? Or is it a recognition that funding for Drug Treatment and Testing Orders hasnot always been adequate, with the result that custodial institutions tend to offer abetter prospect?51

treat-Third, Halliday linked his policy towards persistent offenders with findings onrisk, as set out in Appendix 3 of his report He cites abundant evidence that thebest predictor of reconviction is the number of previous convictions, but fails tonotice that absent from the data is any assessment of the relative seriousness of theindividual offences This is a fault line running through many discussions of risk andcrime prevention The passage refers to offenders who commit a disproportionateamount of crime Is it ‘crime’ in general that policies should be designed to reduce

or, in a system of limited resources, is it not wiser to target offences that are ofparticular concern – notably, the more serious ones? As we shall see, this ambiguityflows through many official statements on persistent offending

Fourth, and despite the rhetoric earlier in the passage, Halliday insisted thatthere should be upper limits (or ceilings) in order to prevent ‘excessive severity’when sentencing offenders with a long record This appears to place his proposalsmidway between progressive loss of mitigation (whose ceilings would, on his view,afford insufficient room for escalation of sentences for repeat offenders) and anunbridled cumulative approach, which he rejected explicitly: ‘a cumulative approachwould lead to disproportionate outcomes As a general principle, the increasedseverity in sentence must retain a defensible relationship with the offences undersentence.’52When he went on to give examples of the scale of enhancement forprevious convictions that he envisaged, it was clear that they were well above existinglevels For example, in relation to domestic burglary and handling stolen goods hepointed out that 14 years was the statutory maximum and that few cases, eventhose of offenders with several previous convictions, were sentenced in the upperquartiles of those ranges

As we shall see below, the Halliday approach has not been followed in all respects

in the new legislation When the Home Office set out its policy in the White Paper,

it had relatively little to say on this subject:

50 ‘Of social variables, drug misuse is most strongly linked with the likelihood of reconviction’: Halliday ( 2001 ), Appendix 3, para 11.

51 Joint Inspection Report ( 2002 ), ch 5 (on the street crime initiative).

52 Halliday ( 2001 ), para 2.15.

Persistent offending should also justify a more severe view and more intensive efforts atpreventing reoffending Increased punishment will be the outcome for those offenderswho have consistently failed to respond to previous sentences We will ensure that such

an outcome is explicit in the statutory framework for sentencing.53

However, an element of confusion appeared in its more general declarations.The most repressive policies were to be directed at ‘serious, dangerous and seriouslypersistent offenders’,54a phrase that contains a crucial ambiguity about whetherthe persistent offenders have committed serious offences, or are merely frequentrepeaters of offences at any level of seriousness Similar confusion runs through theCarter review On the one hand it recognized the problem with the claim about the100,000 persistent offenders,55it urged the greater use of curfews and non-custodialinterventions, it pointed to the strong link between persistent offending and drugproblems, and it recognized that the result of ‘the strict management of drug offend-ers is a risk of rapidly increasing the prison population’.56On the other hand itcriticized the courts for giving ‘greater emphasis to the seriousness of the offencerather than the number of previous convictions’,57referred enthusiastically to the

‘risk assessment of offenders’ without any reference to the seriousness of the offences

‘risked’,58and capped it all by advocating ‘a clear gradient of sentencing severity,which increases with the number of previous offences’ without any reference to theseriousness of the offending.59

Official policy on persistent offenders therefore suffers from ambiguity andinconsistency There is considerable emphasis on risk of reoffending, but oftenwithout distinguishing serious from less serious offences The main feature of theresponse to persistent offending is one of escalating severity, and that assumes thatincreased punitiveness is justified, that defiance makes offences more serious, andthat offenders are rational actors (whereas Home Office evidence shows that theseare often the most disadvantaged people, and that drug addiction plays a majorpart) Little account has been taken of criminological findings on criminal careersand desistance, and on using the desistance research to reorientate the response torecidivism towards a more constructive handling of the relatively few crime-proneyears of many of these offenders.60

6.3.2 The new law

Section 143(2) of the Criminal Justice Act 2003 provides:

In considering the seriousness of an offence (‘the current offence’) committed by anoffender who has one or more previous convictions, the court must treat each previousconviction as an aggravating factor if (in the case of that conviction) the court considersthat it can reasonably be so treated having regard, in particular, to –

53 Home Office ( 2002 ), para 5.10 54 Home Office ( 2002 ), para 5.7.

55 Carter ( 2003 ), p 16 56 Carter ( 2003 ), p 29 57 Carter ( 2003 ), p 18.

58 Carter ( 2003 ), pp 27–8 59 Carter ( 2003 ), p 29.

60 See now Bottoms, Shapland et al ( 2004 ).

(a)the nature of the offence to which the conviction relates and its relevance to thecurrent offence, and

(b)the time that has elapsed since the conviction

The drafting of this provision indicates that a court is bound to treat each previousconviction as a factor that aggravates the current offence, and must do so each timethe offender is sentenced.61Thus, if a person who was sentenced in April 2005 for

an offence, and had five previous offences that were taken to aggravate the sentence,were to come up for sentence again in May 2006, the court would be required totreat each of his (now) six previous convictions as aggravating the sentence

It may be argued that the force of the mandatory words, ‘must treat as anaggravating factor’, is considerably softened by the later clause, ‘if the court considersthat it can reasonably be so treated’ But a more straightforward reading would bethat the court’s decision on whether a previous conviction ‘can’ (should?) be treated

as aggravating is to be determined by its assessment of its relevance and recency,

as set out in (a) and (b), and not by any broader views about whether previousconvictions should be allowed to increase sentence The Explanatory Notes to theAct state baldly that recent and relevant convictions ‘should be regarded as anaggravating factor which should increase the severity of the sentence’

What should be the criterion of whether prior convictions are relevant to thecurrent offence? The Halliday report noted that most persistent offenders have amixed criminal record, and therefore argued that ‘less weight should be given towhether previous and current offences are in the same category’, so that ‘the keypoint is whether the previous offences justify a more severe view’.62Insofar as thissuggests that the seriousness of the previous offences is the crucial issue, it seemsdifficult to reconcile with reference in s 143(2) to relevance, which may be taken

to indicate a similarity of subject matter In the past, the practice in sentencingfor offences of violence has been for courts to pay more attention to previousconvictions for offences of violence than to others; and the same might be said ofsexual cases.63Where there is a record of offences of dishonesty or burglary, courtsmay decide to treat the person as a ‘professional’ (see part6.4below) However,both Appendix 3 of the Halliday report and David Farrington’s review of criminalcareer research lead to the conclusion that the typical pattern is a small degree ofspecialization ‘superimposed on a great deal of generality or versatility in offending’,and that the majority of offences of violent offenders are non-violent.64This raisesthe question whether mere similarity of offence category, rather than targetingparticular types of victim, should be relevant at all It could also be argued thatthe notion of lapse, underlying the concession of first offenders, should not be

61 Convictions from other jurisdictions may also be taken into account (subsections (4) and (5)), but it is unclear whether a conviction followed by a discharge counts as a previous conviction in this context.

62 Halliday ( 2001 ), para 2.17; later in the same paragraph, however, the report states that ‘completely disparate previous convictions should be given less weight’.

63 See Wasik ( 1987 ), pp 108–9 64 Farrington ( 1997 ), p 380.

applied simply to the heterogeneous category of ‘crime’ Human weakness in losingone’s temper momentarily and punching another may be regarded as differentfrom human frailty in succumbing to the temptation of economic crime There istherefore some ground for arguing that a first offence of violence by someone withprevious property convictions should be treated as ‘out of character’, and should

be mitigated to some extent But this argument cannot be pressed too far It would

be absurd to imply that everyone is entitled to one ‘discounted’ crime of violence,one ‘discounted’ fraud, one ‘discounted’ sexual offence and so on Thus Hallidaysuggested that the seriousness of the offences is the primary determinant, in terms

of ‘whether there is a continuing course of criminal conduct’.65Section 143(2)(a)seems to have a different emphasis, and a narrower notion of relevance – definedaccording to types or categories of offending – may take root

There is widespread agreement that a gap in offending should be taken to ish the effect of previous convictions, and s 143(2)(b) accordingly requires courts tohave regard to the time that has elapsed since each previous conviction An example

dimin-of this from the common law is Fox (1980),66where the Court of Appeal reducedthe sentence on a man aged 35 convicted of grievous bodily harm who had twoprevious convictions many years earlier: ‘In our judgment, his previous record ofviolence when he was in his late teens and mid-twenties should have been left out ofaccount in deciding what action to take.’ Various justifications may be offered forthis concession – for example the offender deserves credit for going straight, or thepresent offence is to some extent ‘out of character’ in terms of his recent behaviour,

or the conviction-free gap makes it less likely that he will reoffend – but the moststraightforward approach is to affirm the underlying principle of the Rehabilitation

of Offenders Act 1974 Generally speaking, it is unnecessarily harsh if a person has

to bear the burden of previous convictions indefinitely: after a number of years aperson should be able to regain full rights as a citizen, and such a principle mayeven provide an incentive not to reoffend Many US guideline systems provide forthe ‘decay’ of previous convictions after 10 years, and this has been adopted, forexample, in the proposed South African sentencing code.67

What is missing from s 143(2) is any reference to an overall proportionalityconstraint, as recommended by the Council of Europe and adopted, for example, inSweden.68The wording of s 143(2) seems consistent with a cumulative approachthat increases the severity of the sentence on each subsequent conviction, allowingsentences (e.g for theft from shops) to ascend to two, two-and-a-half or three years’

65 Halliday ( 2001 ), para 2.17.

66 (1980) 2 Cr App R (S) 188; see also Bleasdale (1984) 6 Cr App R (S) 177 (four years without trouble

for a man of 22 ‘is an important feature in his favour’).

67 South African Law Reform Commission ( 2000 ), s 42: ‘where a period of 10 years has passed from the date of completion of the last sentence and the date of commission of any subsequent offence the last conviction and all convictions prior to that must be disregarded for the purposes

of sentencing’.

68 Von Hirsch and Jareborg ( 1989 ): see ch 29.4, ‘the court shall to a reasonable extent take the offender’s previous criminality into account’.

imprisonment We have seen that Halliday insisted on a proportionality constraint.Section 143(1) may be taken to indicate a general principle of proportionality insentencing but, as we saw in Chapter3.5above, s 142 requires courts to have regard

to other purposes of sentencing such as deterrence and public protection – whichmay be thought to authorize long sentences for persistent offenders However, inParliament it was said that s 143(2) was not intended to lead to disproportionatesentences,69and it therefore falls to the Sentencing Guidelines Council to ensurethat this intention is carried through into the relevant guidelines

Finally, we should return to the mandatory wording of s 143(2) Does the tion that ‘the court must treat each previous conviction as an aggravating factor’mean that Parliament intended to rule out the use of community sentences forrepeat offenders? At one level it does not: a first offence may warrant a fine, andthe two or three subsequent offences may justify only community sentences Thequestion then is whether the offence – say, theft from shops – is so serious that acustodial sentence can be justified for a repeat offender If the answer is yes,70then

injunc-it may seem inconsistent winjunc-ith the principle of treating each previous conviction asaggravating if a court is to decide on a community sentence For example, some suchoffenders commit frequent thefts of items such as toiletries, with a view to sellingthem in order to raise money to buy drugs.71The offender may come before thecourt with 30, 40 or more previous convictions If the court wishes to tackle what

it regards as the underlying cause of offending (drug-taking) by making a nity order, would it be lawful to do so? Halliday himself wanted to see communitysentences used more widely in such cases, but there was always a conflict with hisproposed policy on persistent offending,72and the mandatory wording of s 143(2)heightens the conflict On one view a court would be acting unlawfully if it dealtwith an offender, whose record disclosed many recent and relevant convictions, bymeans of a community sentence: each previous conviction would be an aggravatingfactor, and the starting point for theft would therefore be a significant custodialsentence Only if there were a guideline that placed a low ceiling on sentences forrepeated non-serious offending would there be a possibility of imposing a commu-nity sentence And even then, would not the court have to demonstrate compliancewith s 143(2) by saying that the previous convictions do aggravate the sentence butthat it still falls beneath the custody threshold?73

commu-69 Baroness Scotland, HL Deb 24 Feb 2003.

70 See Page et al., unreported, 8 Dec 2004, where Rose LJ held that persistent shop theft on a

minor scale might justify one month’s imprisonment; a history of persistent similar offending on

a significant scale ‘would often merit no more than 12 to 18 months’; but where offences were attributable to drug addiction, a community sentence aimed at tackling that addiction would often

be appropriate See further ch 10.6 below.

71 See the instructive discussion of such a case by Jones ( 2002 ).

72 Jones ( 2002 ), pp 185–6, citing Halliday ( 2001 ), para 6.6.

73 Cf cases where drug treatment and testing orders were imposed for offences serious enough to

attract up to three years’ custody, e.g Kelly [2003] 1 Cr App R (S) 472 and Belli [2004] 1 Cr App

R (S) 490, and below, ch 10.6.3.10.

6.3.3 The probable outcomes

Having considered the policy towards persistent offenders generally, and the specificterms of s 143(2), we now turn to consider the probable effects of the new law.Three spheres of impact will be assessed – sentencing practice, the prison population,and public protection

1 Sentencing practice Halliday envisaged that the principal constraint on

sen-tencers would be definitive guidelines from the Sentencing Guidelines Council,which he expected to identify different starting points according to the weight ofthe previous criminal record The only current guidelines with this degree of detail

are those for domestic burglary, handed down by the Court of Appeal in McInerney

and Keating (2003)74before the 2003 Act was passed These establish distinct ing points for first-time, second-time and third-time domestic burglars committing

start-a ‘ststart-andstart-ard’ burglstart-ary, start-and start-are therefore start-along the lines suggested by Hstart-allidstart-ay Thisapproach was taken because of the minimum sentence for the third domestic bur-glary (see part6.7below), and the belief that previous record is more important

in burglary (presumably because it has a relatively high proportion of specialists,unlike many other crimes) But how might a court, applying the 2003 Act, deal with

a case such as Woods (1998)?75A vicar returned to his vicarage to find the offenderdozing on the floor, having broken in and put various items from the house in abag ready to take with him The offender had 121 previous convictions for offences

of burglary and theft, not of great seriousness, and had been released from prisononly 21 days before the burglary The trial judge sentenced him to six years’ impris-onment, saying that there was no hope of rehabilitating or deterring the offenderand that therefore an incapacitative sentence was needed in order to protect thepublic The Court of Appeal acceded to the proposition that the offender was not aprofessional but an incompetent yet frequent opportunist, usually looking for food

or money The Court reduced the sentence from six to four years This was still awholly disproportionate sentence for a burglary that is fairly low down the scale.Under the 2003 Act a court would be bound to find that his previous convictionswere both recent and relevant, but the sentencing guidelines (underpinned by theminimum sentence for the third domestic burglary) would indicate a starting point

of at least three years If a court were really to regard each of his 121 convictions asaggravating the current offence, then it might well impose a sentence even longer

than the four years upheld in Woods There is nothing in s 143(2) to direct courts to

keep sentences for minor offences low, and it will fall to the Sentencing GuidelinesCouncil to reinforce what Baroness Scotland stated in the parliamentary debates.One of the curious features of the pre-2003 law was that the Court of Appealfailed to analyse (and only rarely referred to)76the statutory provision on persistentoffenders That is unlikely to happen under the 2003 Act, and indeed the Court ofAppeal may be called upon to interpret s 143(2) before the Council has laid down

74 [2003] 2 Cr App R (S) 240, discussed above, ch 4.4.11 75 [1998] 2 Cr App R (S) 237.

76 A rare example was Spencer and Carby (1995) 16 Cr App R (S) 482 at pp 485–6.