Sentencing aims, principles and policies

CHAPTER 3
Sentencing aims, principles and policies
3.1 The aims of the criminal justice system
The ‘criminal justice system’ is not a structure which has been planned as a system.
Norisitsoorganized that the several interlocking parts operate harmoniously. In
England and Wales, as in many other jurisdictions, the administration of criminal
justice has grown in a piecemeal way over the years, with separate phases of devel-
opment leaving their mark. To refer to a ‘system’ is therefore merely a convenience
and an aspiration. It should not be assumed that the various arrangements were
planned or actually operate as a system, although it remains necessary to recog-
nize the interdependence of the different parts and to incorporate this into any
planning.
It is important to distinguish the aims of the criminal justice system from the aims
of sentencing, which merely relate to one element. The system encompasses a whole
series of stages and decisions, from the initial investigation of crime, through the
various pre-trial processes, the provisions of the criminal law, the trial, the forms
of punishment, and then post-sentence decisions concerned with, for example,
supervision, release from custody and recall procedures. It would hardly be possible
to formulate a single meaningful ‘aim of the criminal justice system’ which applied
to every stage. It is true that one might gather together a cluster of aims: for example,
the prevention of crime, the fair treatment of suspects and defendants, due respect
for the victims of crime, the fair labelling of offences according to their relative
gravity and so on. But to combine these into some overarching aim such as ‘the
maintenance of a peaceful society through fair and just laws and procedures’ is
surely to descend into vacuity, since it gives no hint of the conflicts that arise and
the priorities that need to be determined. The Home Office’s first Statement of
Purpose reveals the conflicts but fails to indicate priorities:
To work with individuals and communities to build a safe, just and tolerant society
enhancing opportunities for all and in which rights and responsibilities go hand in
hand, and the protection and security of the public are maintained and enhanced.
1

1 www.homeoffice.gov.uk.
67
68 Sentencing aims, principles and policies
This generalized purpose does not recognize that different stages may have their
distinct aims and purposes. It needs supplementing with more focused aims, but
those enumerated by the Home Office do not carry the issue much further. These
are
(i) to reduce crime and the fear of crime, tackle youth crime and violent, sexual and
drug-related crime, anti-social behaviour and disorder, increasing safety in the home
and public spaces;
(ii) to reduce organized and international crime, including trafficking in drugs, peo-
ple and weapons, and to combat terrorism and other threats to national security,
in co-operation with European Union (EU) partners and the wider international
community;
(iii) to ensure the effective delivery of justice, avoiding unnecessary delay, through efficient
investigation, detection, prosecution, trial and court procedures. To minimize the
threat to and intimidation of witnesses and to engage with and support victims; and
(iv) to deliver effective custodial and community sentences to reduce reoffending and
protect the public, through the prison and probation services, in partnership with
the Youth Justice Board.
These aims are undoubtedly important, even if the attempt to highlight some
forms of crime results in leaving out others. However, once again, there is no
acknowledgment of the inevitable conflicts, no reference to human rights, and no
reference to appropriate international documents (e.g. European Convention on
Human Rights, United Nations Convention on the Rights of the Child).
Unrealistic aims should not be set for individual decisions in the criminal justice
system. We saw earlier
2
that only a small proportion of crimes come before the
courts for a sentencing decision – around 2 per cent on Home Office figures. Even

granted that publicity may make it appear that the courts are dealing with a higher
proportion than this, the potential of sentencing for altering the frequency and
patterns of offending in society is severely handicapped by the fact that relatively
few offences result in the passing of a sentence. However, it may be assumed that
sentencing fulfils an indispensable public function within the criminal justice sys-
tem: without the panoply of police, penal agents and courts, there would surely be
more crime. There is at least some evidence that law and order would break down
in the absence of police, for example.
3
Butitdoes not follow from any of this that
increases in sentence levels will bring about increases in general crime prevention,
as we shall see in the discussion of deterrence theory in part 3.3.2 below.
The conscientious pursuit of crime preventionis, however, a worthy objective of a
criminal justice system as a whole, and considerable developments have taken place.
Since at least the early 1980s the Home Office has devoted considerable attention
2Inch.1.4 above.
3Evidencefor this might be derived from the spread of lawbreaking, mostly property offences, during
the police strikes in Melbourne in 1918 and Liverpool in 1919, and during the immobilization of
the Danish police force in 1944. It is argued by Mathiesen (1990), pp. 62–3, that these were such
atypical situations that they leave the propositions in the text as unsupported assertions.
3.1 The aims of the criminal justice system 69
to different forms of crime prevention, in order to identify and to carry forward
the most effective methods of crime prevention. Some of these begin with fam-
ily planning and parenting, through pre-school facilities to the identification and
monitoring of children ‘at risk’ of offending.
4
Then there is situational crime pre-
vention, which the Home Office has long encouraged through a variety of initiatives,
such as altering the designs of buildings or vehicles in order to reduce the oppor-
tunity for certain kinds of crime. This ‘target-hardening’ approach has been used

to increase surveillance (e.g. the now widespread use of cameras in public places,
on public transport and in shopping centres), to make houses more secure against
burglars, and so forth.
5
A further possibility is social crime prevention, although in
the present government’s policy this is often termed ‘community crime prevention’
and includes ‘zero tolerance’ approaches to incivilities as well as improvements to
housing, social and recreational facilities, education and employment.
6
Te c hniques
of policing may also offer possibilities for crime prevention, although the prospects
of success are often grossly overestimated by references to ‘more police on the beat’
as a solution to alleged increases in the crime rate. There is, however, evidence that
in some circumstances certain techniques of policing can bring crime prevention
benefits.
7
There is much promise in some crime prevention strategies, insofar as they are
shown to reduce crime and thereby reduce the load on the law enforcement agencies
and the labelling of people as offenders. The history of ‘auto-crime’ shows the
considerable impact of introducing steering locks in the 1960s in reducing thefts and
takings of cars–afar more significant reduction than could have been achieved by all
but the most draconian sentencing policy – and in the 1990s motor manufacturers
co-operated in improving car security as part of a renewed effort against these types
of crime. However, although some crime prevention strategies appear so promising
that they should be pursued with much greater vigour than at present, there are
at least three drawbacks which must be borne in mind. One is that the number
of small local projects far outstrips the amount of careful and rigorous evaluation.
Schemes are often difficult to evaluate, and not just because one has to investigate
possible ‘displacement’ effects, in the shape of lawbreaking of other kinds or in
other areas. The political attractions of crime prevention initiatives are sometimes

allowed to run ahead of proper assessments of their effectiveness.
8
Asecond danger
is that the schemes will be used to spread the net of social control, promoting so-
called ‘community’ initiatives in a waywhich increases state control over individuals,
families and neighbourhoods and therefore brings other disadvantages. Insufficient
attention has been paid to ethical issues in crime prevention, raised by a number
of techniques (such as CCTV). A third unwelcome consequence is that situational
approaches might conduce to the mentality of a ‘fortress’ society, surrounded by
locks, bars and unbreakable articles. This might heighten fear of crime, even if it
reduces objective risk. Despite these drawbacks, it remains the best policy to try
4Graham (1998). 5 Ekblom (1998), Pease (1998). 6 Hope (1998).
7Jordan (1998). 8 For an overview and critical discussion, see Bottoms (1990).
70 Sentencing aims, principles and policies
to prevent crime before it occurs, so long as this can be achieved within a rights-
based framework. However, when the government abandoned its much-trumpeted
10-year Crime Reduction Programme in 2002, after only three years, it was evident
that the main source of disappointment stemmed from setting over-ambitious
targets too quickly, without proper monitoring and evaluation.
9
If prevention does not work, then the state must be prepared to respond to an
offence that has been committed. The immediate danger is that sentencing will
be expected to function efficiently as a crime prevention mechanism, when there
are well-documented reasons why this may not happen. Two clear reasons why
sentencing and crime rates may vary independently are (i) that crime rates are
affected by demographic factors such as the age profile of the population and by
changes in the availability of desirable and stealable goods (such as mobile phones);
(ii) that fewer than half of all crimes are reported to the police, as we saw in Chapter
1.4 above.
10

When there is a formal response to an offence this does not always
mean prosecution–conviction–sentence, since, as we saw in Chapter 1.4, there are
various methods of diversion available. For those cases that are brought to court,
however, sentencing is a process that has considerable social significance in its own
right. Conviction involves the public labelling of people as offenders. The sentencing
decision can often be seen as the core of the labelling or censuring process by giving
ajudgment of ‘how bad’ the offence was, and by translating that judgment into the
particular penal currency of this country at this time. Sentencing has an expressive
function and, as Durkheim argued, ‘the best punishment is that which puts the
blame . . . in the most expressive but least costly form possible’.
11
This expressive or censuring function is carried out by means of imposing coer-
cive measures on convicted offenders. The imposition of punishment requires justi-
fication. We should not be satisfied with the proposition that anyone who commits
any offence forfeits all rights, and may be dealt with by the state in whatever manner
the courts decree. That would be to suggest that any convicted person is at the dis-
posal of the criminal justice system, and has no relevant rights. Instead, we should
seek strong justifications for contemporary sentencing practices, not least because of
the increasing use of imprisonment and the greater restrictiveness of non-custodial
sentences in many countries. But before turning to consider the possible rationales
for sentencing, it is first necessary to say something about the institution of state
punishment.
3.2 Justifying state punishment
Whence does the state acquire its right to punish, and what sustains it? A proper
answer to these questions would require a substantial foray into political philosophy.
9Maguire (2004). 10 See further Bottoms (2004), pp. 60–1.
11 Quoted in Garland (1990), p. 46.
3.2 Justifying state punishment 71
All that can be done here is to sketch some of the lines of justification.
12

It is often
assumed that the right to punish is simply one aspect of the modern sovereign
state, but any such assumption is disputed by those who proclaim that victims and
their families, or victims and communities (through restorative justice), ought to
be central to responses to crime.
13
Justifications for assigning the central role to the state are often derived from
social contract theories, the essence of which is that citizens give up their ‘natural’
righttouse force against those who attack their interests and hand it over to the
state, in return for the state’s promise to protect them by maintaining law and
order.
14
Citizens retain a limited right of self-defence, but apart from that the state
takes charge of enforcing the law, maintaining courts and providing the institutions
of punishment. Without some such idea of contract, the co-operation on which
society rests could not be attained, it is argued. The state then has the responsibility
of ensuring peaceable co-operation, and one aspect of that is to establish a category
of wrongs that amount to crimes. It is the state’s task to provide police, prosecutors
and courts to respond to these wrongs. Individual victims may bring civil actions
against the perpetrators, but it is in principle for the state to prosecute and (on
conviction) to provide the institutions of sentencing. Another approach would
be to justify the state’s role in punishment by reference to the need to displace
individual revenge and retaliation by maintaining a social practice that constitutes
an independent and authoritative response to crime.
15
This does not constitute
the state as a ‘proxy retaliator’: the state has the duty to act with justice and with
humanity in discharging the function of punishment, and often there may be a
‘displacement gap’ between what the public or the media would like to see by way of
punishment, and what the state’s institutions can and should provide. Regulating

that gap and its social consequences is one of the modern state’s more difficult
obligations.
The importance of punishment being in the hands of state institutions rather than
victims or other individuals resides in rule-of-law values. Decisions on punishment
should be taken by an independent and impartial tribunal, not by individuals with
an emotional involvement in the events. The outcome should not be dependent on
whether the victim is vengefulor forgiving, but should be dependent on the impartial
application of settled principles, notably principles that recognize the offender as
a citizen capable of choice and that regard proportionality of sentence to offence
as a key value.
16
The state therefore has the role of providing the institutions for
an authoritative response to wrongs, which constitute a public valuation of the
offender’s conduct.
17
Sometimes these notions are expressed in terms of the state
and its courts being more ‘objective’ than victims and their families, but one must
12 For accessible discussions, see MacCormick and Garland (1998), Gardner (1998) and Duff (2001).
13 E.g. Christie (1977).
14 See MacCormick and Garland (1998) for discussion and variations on this theme.
15 Gardner (1998). 16 Ashworth (2002).
17 See further von Hirsch and Ashworth (2005), ch 2.
72 Sentencing aims, principles and policies
beware of the concept of objectivity here. Issues of crime and punishment have
become intensely political in recent years and, even if sentences are objective in
the sense that they are not chosen by victims or their representatives, they are not
objective in the sense of being free from the political posturing or vote-catching
policies that have tended to shape sentencing legislation (and therefore judicial
sentencing) in recent decades.
Thus whether one takes the justification for state punishment to be an aspect

of the idea of a social contract, or (more pragmatically) to be the carrying out of
a displacement function that is essential to social co-operation, there are problems
in translating the justification to any particular criminal justice system. There are
many signs of what David Garland has termed ‘the decline of the sovereign State’,
18
and, even if some of his analysis is less compelling than it might appear,
19
it is
surely true that the simple model in which the state provides for the security of its
subjects is not sustainable in many countries. Responsibility is being devolved to
private entrepreneurs and to local authorities, and crime is perceived as a major
social problem still. At some times in some countries, the legitimacy of the state and
its institutions suffers collapse, and those dire circumstances would force recon-
sideration of the basic principles.
20
Thus we might conclude with Antony Duff
that, although there may be justifications for the state taking responsibility for
criminal justice, they are contingent on the state fulfilling its side of the agree-
ment,
21
and in many countries that is in doubt. This area of doubt makes it all
the more important to scrutinize the justifications for sentencing policy in general,
for the types of sentence that are used, and for the conditions that they impose on
offenders.
3.3 The rationales of sentencing
3.3.1 The argument for declaring a primary rationale
When judges are discussing sentencing, one of the most frequent topics is discretion.
Some of the constitutional dimensions were mentioned in Chapter 2, but another
dimension is the constant tension between flexibility and the rule of law. There
are many who would agree that sentencers ought to have sufficient discretion to

take account of the peculiar facts of individual cases. So be it. But does that remove
the argument for bringing the rule of law as far into sentencing decisions as possible?
The rule of law, in this context, means that judicial decisions should be taken
openly and by reference to standards declared in advance.
22
It is one thing to agree
that judges should be left with discretion, so they may adjust the sentence to fit
the particular combination of facts in an individual case. It is quite another to
suggest that judges should be free to choose what rationale of sentencing to adopt
in particular cases or types of case. Freedom to select from among the various
18 Garland (2000). 19 Zedner (2002). 20 For references, see Ashworth (2002b), pp. 580–1.
21 Duff (2001), p. 197. 22 Raz (1979), ch. 11.
3.3 The rationales of sentencing 73
rationales is a freedom to determine policy, not a freedom to respond to unusual
combinations of facts. It is more of a licence to judges to pursue their own penal
philosophies than an encouragement to respond sensitively to the facts of each case.
It is fairly well established that a major source of disparity in sentencing is
the difference in penal philosophies among judges and magistrates.
23
Yet many
judges and magistrates place great importance on the freedom to pursue whatever
approach they think appropriate ‘on the facts of the case’.
24
One notable decision
of the Supreme Court of Victoria expresses what many judges may believe:
The purposes of punishment are manifold and each element will assume a different
significance not only in different crimes but in the individual commission of each
crime . . . Ultimately every sentence imposed represents a sentencing judge’s instinctive
synthesis of all the various aspects involved in the punitive process.
25

The inscrutable idea of an ‘instinctive synthesis’ comes close to another notion,
which is that the various aims of sentencing should be ‘balanced’ in each case.
Indeed, the Sentencing Reform Act of 1984 in the United States required the US
Sentencing Commission to devise guidelines that reflected proportionality, deter-
rence, public protection and offenders’ treatment needs – aims that were listed
without recognition that they conflict, and that priorities must be established.
If there is thought to be some value in each of these purposes, what should be
done?
It is often assumed that there are only two alternative courses: either (i) to declare
a single rationale, or (ii) to allow sentencers a fairly free choice among several
rationales. Critics of the first approach argue that it is too rigid, especially when
there is such a wide range of crimes and criminals. They may then assume that the
second approach is the only ‘realistic’ one. They may argue that the second approach
is more ‘balanced’ or is ‘multi-faceted’, thereby contrasting its practicality with the
academic, even ascetic regime of a single rationale. But there is a third possibility,
which is both practical and consistent with the rule of law: (iii) to declare a primary
rationale, and to provide that in certain types of case one or another rationale might
be given priority. This approach has been operating in Sweden since 1989, with
desert or proportionality as the primary rationale and other aims having priority
in certain types of case.
26
It was also the approach embodied in the Criminal Justice
Act 1991, with desert as the primary rationale and incapacitation having priority
in certain types of case. And it received the approval of the Council of Europe in its
recommendation on ‘Consistency in Sentencing’:
23 See Hogarth (1971), cited in ch. 1.6 above, and the wider review of research by the Canadian
Sentencing Commission (1987), para. 4.1.2.
24 See ch. 1.6 above on this concept.
25 Williscroft [1975] VR 292, at pp. 299–300; see also Young [1990] VR 951.
26 For the text of the law in English, see von Hirsch and Jareborg (1989); for discussion, see Jareborg

(1995).
74 Sentencing aims, principles and policies
A.1 The legislator, or other competent authorities where constitutional principles and
legal traditions so allow, should endeavour to declare the rationales for sentencing.
A.2 Where necessary, and in particular where different rationales may be in conflict,
indications should be given of ways of establishing possible priorities in the application
of such rationales for sentencing.
A.3 Where possible, and in particular for certain classes of offences or offenders, a
primary rationale should be declared.
27
However, the government appears not to regard itself as bound by this kind
of clearly structured approach. The scheme of the 1991 Act has been abandoned,
and in its place we have a law that seems to embody the worst of ‘pick-and-mix’
sentencing. Section 142 of the Criminal Justice Act 2003 provides:
Anycourt dealing with an offender in respect of his offence must have regard to the
following purposes of sentencing –
(a) the punishment of offenders,
(b) the reduction of crime (including its reduction by deterrence),
(c) the reform and rehabilitation of offenders,
(d) the protection of the public, and
(e) the making of reparation by offenders to persons affected by their offences.
This invites inconsistency, by requiring judges to consider a variety of different
purposes and then, presumably, to give priority to one. However, it seems possible
that its effect will be blunted by another provision in the 2003 Act, which the Sen-
tencing Guidelines Council has adopted as the touchstone for its guidelines. Thus,
having set out the terms of s. 142, the Council goes on to state that ‘the sentencer
must start by considering the seriousness of the offence’, and then quotes s. 143(1):
In considering the seriousness of any offence, the court must consider the offender’s
culpability in committing the offence and any harm which the offence caused, was
intended to cause or might foreseeably have caused.

The remainder of the Council’s guideline on Overarching Principles focuses on
the proportionality principle in s. 143, without returning to s. 142, and makes it
clear that s. 143 will underpin the guidelines it issues.
28
It remains to be seen how
closely the guidelines are followed, and what happens to any judge or magistrate
who purports to ‘have regard to’ one of the purposes in s. 142 rather than to the
guidelines.
29
The enactment of s. 142 makes it all the more important to examine six contem-
porary rationales of sentencing: deterrence,rehabilitation, incapacitation, desert,
27 Council of Europe (1993), p. 6. 28 SGC, Overarching Principles – Seriousness (2004).
29 It should be mentioned that s. 142 does not apply to the sentencing of offenders under 18. However,
there are also conflicting rationales in respect of them – s. 37 of the Crime and Disorder Act 1998
states that the aim should be ‘to prevent offending by children and young persons’, whereas s. 44
of the Children and Young Persons Act 1933 enjoins courts to ‘have regard to the welfare of the
child or young person’. See further ch. 12.1 below.
3.3 The rationales of sentencing 75
social theories, and reparation or restoration. Each of these aims has a considerable
philosophical background and penological context, which cannot be set out in full
here. Readers are referred to a recent anthology of readings, with commentary and
bibliography, for further study.
30
3.3.2 Deterrence
31
Deterrence is one of several rationales of punishment which may be described
as ‘consequentialist’, in the sense that it looks to the preventive consequences of
sentences. In fact, deterrence is merely one possible method of producing crime
prevention through sentencing: it relies on threats and fear, whereas rehabilitation
and incapacitation adopt different methods of trying to achieve a similar end, as

we shall see below. It is important to draw the distinction between individual (or
special) deterrence and general deterrence. The latter aims at deterring other people
from committing this kind of offence, whereas individual deterrence is concerned
with deterring this particular person from reoffending. A system which regards
individual deterrence as the main goal would presumably escalate sentences for
persistent offenders, on the reasoning that if non-custodial penalties fail to deter
then custody must be tried, and if one year’s custody fails to deter, two years must
be tried, and so on. It is not the gravity of the crime but the propensity to reoffend
which should be the main determinant of the sentence. Although this approach
seems to underlie the latest provision on persistent offenders,
32
it is rarely adopted
as the primary rationale of a sentencing system.
More significant is general deterrence. Jeremy Bentham was its chief proponent,
and he started from the position that all punishment is pain and should therefore
be avoided. However, punishment might be justified if the benefits (in terms of
general deterrence) would outweigh the pain inflicted on the offender punished,
and if the same benefits could not be achieved by non-punitive methods. Sentences
should therefore be calculated to be sufficient to deter others from committing this
kind of offence, no more and no less. The assumption is that citizens are rational
beings, who will adjust their conduct according to the disincentives provided by
sentencing law. The same assumption leads to a belief in marginal deterrence – that
increasing penalty levels by a certain amount will result in a decline in offending.
Modern economic theorists such as Richard Posner adopt a similar approach, view-
ing punishments as a kind of pricing system.
33
Less sweeping is the rational choice
perspective, adopted by criminologists such as Ronald Clarke as an explanation of
certain types of offending and used to generate specific preventive strategies. The
argument is that particular types of crime tend to result from a form of rational

calculation (usually termed ‘bounded rationality’), and that the responses to such
crimes should take account of this and combat it.
34
30 Von Hirsch and Ashworth (1998). 31 Von Hirsch and Ashworth (1998), ch. 2.
32 S. 143(2) of the 2003 Act, analyzed in ch. 6.3.2 below.
33 Posner (1985), excerpted in von Hirsch and Ashworth (1998), ch. 2; see also Pyle (1995).
34 Cornish and Clarke (1986).
76 Sentencing aims, principles and policies
Criticisms of deterrence theory may be divided into the empirical and the prin-
cipled. The main empirical criticism is that the factual data on which a deterrent
system must be founded do not exist. Reliable findings about the marginal general
deterrent effects of various types and levels of penalty for various crimes are hard
to find. For example, sophisticated techniques have been applied in attempts to
assess the deterrent efficacy of the death penalty, without yielding clear and reliable
results.
35
Anecessary element in research is a proper definition of deterrence, to
establish that fear of the legal penalty was the particular factor that led to avoidance
of the proscribed conduct. Deterrence must operate (if at all) through the potential
offenders’ minds, so it is essential that they know about the severity of the probable
sentence, take this into account when deciding whether to offend, believe that there
is a non-negligible risk of being caught, believe that the penalty will be applied to
them if caught and sentenced, and refrain from offending for these reasons.
36
These
subjective beliefs are vital components in the operation of deterrent policies, and all
must therefore be investigated if research is to be reliable. Few studies satisfy these
criteria, and they provide no basis for sentencing policies that involve increasing
severity in order to reduce offending levels. This was the major finding of the Cam-
bridge study, commissioned by the Home Office, although it did find that there

was better evidence of the deterrent effect of a (believed) high risk of detection
than of (believed) penalties.
37
The Halliday report reviewed the evidence and also
concluded that the limited evidence ‘provides no basis for making a causal connec-
tion between variations in sentence severity and differences in deterrent effects’.
38
Asubsequent international review by Doob and Webster recognized the intuitive
attraction of the deterrent hypothesis but still found that the evidence indicated
‘that sentence severity has no effect on the level of crime in society’.
39
There is a little research which suggests that certain forms of offence which
tend to be committed by people who plan and think ahead may be susceptible to
deterrent sentencing strategies: Richard Harding, for example, found that robbers
tended to desist from arming themselves with guns if there was a significant extra
penalty for carrying a firearm.
40
This may be taken to bear out the proposition
that general deterrence is more likely to be effective for planned or ‘professional’
than for impulsive crimes, although Harding argues that deterrent sentences need
to be combined with publicity and appropriate ‘social learning’ opportunities if
they are to have significant preventive effects. A counterpoint is provided by David
Riley’s study of drink drivers, in which he shows that the problems of a general
deterrent strategy lie in drivers’ optimism about the risk of being caught, ignorance
of the penalty, and ignorance of the amount of alcohol consumption needed to
commit an offence.
41
Further studies have examined the potential deterrent effect
of increased enforcement by the police, but it seems that a general crime prevention
35 Hood (2002), ch. 7. 36 See Bottoms (2004), p. 65.

37 Von Hirsch et al. (1999), chs. 3 and 7. 38 Halliday (2001), p. 129.
39 Doob and Webster (2003), p. 143. 40 Harding (1990). 41 Riley (1985).
3.3 The rationales of sentencing 77
strategy with publicity and attempts to change people’s attitudes is likely to be more
effective than either sentencing or enforcement changes alone.
42
Another area in
which the potential for legal deterrence appears not to be great is burglary: inter-
views with burglars suggest that most of them are not rational calculators but rather
short-term hedonists or eternal optimists.
43
Particularly interesting is the finding
of Ros Burnett and Shadd Maruna that, although the majority of their convicted
prisoners wanted to desist from crime after their release, only a minority succeeded
in doing so and it tended to be a philosophy of hope that distinguished them. The
notion of austere prison conditions as a deterrent was simply not enough.
44
This
dearth of supporting evidence leaves some authors undaunted, since they argue that
‘commonsense reasoning about general prevention’ can be used instead.
45
There is
apoint here: general deterrence can indeed work, given the necessary favourable
circumstances.
46
But the available research surely demonstrates the danger of gen-
eralizing from intuitions, or one’s personal experience, to the probable reactions of
others.
47
Reliable and precise evidence is required, and it is not available.

Principled criticisms of deterrence theory would apply whether or not there is
satisfactory evidence of general deterrent effects. One such criticism is that the
theory could justify the punishment of an innocent person if that were certain
to deter several others: a simple utilitarian calculus would allow this to happen,
without any respect for the rights of the innocent person. Another, more realistic
criticism is that the theory can justify the imposition of a disproportionately harsh
sentence on one offender in order to deter several others from committing a similar
offence. This is the so-called ‘exemplary sentence’. English judges have passed such
sentences from time to time,
48
and some would argue that such decisions have been
the product of political or ‘media’ pressure to respond to public anxiety about a
certain type of crime. One incident which has become part of judicial lore is the
passing of exemplary sentences on certain offenders after the Notting Hill race riots
in 1958. It is argued that such sentences may be justified by the consequences, which
in this case were reductions in racial troubles in Notting Hill (although there were
similar troubles in other cities in the following months). But who can assert that it
was the exemplary sentences which caused the reduction in the number of offences
which otherwise would have taken place? Might it not be the case that the police
had arrested and charged the ringleaders, and without them there would be no
continuation? Or that increased police patrols dramatically increased the perceived
risk of being caught? The Notting Hill case serves only to emphasize the formidable
difficulties of gathering evidence on the effectiveness of exemplary sentences as
short-term deterrents.There must be no other plausible explanations for the changes
in people’s behaviour: otherwise, one cannot be confident of interpreting a sequence
of social events correctly.
42 Riley (1991). 43 Bennett and Wright (1984), chs. 5 and 6.
44 Burnett and Maruna (2004). 45 E.g. the classic book by Andenaes (1974).
46 Nagin (1998). 47 Mathiesen (1990), pp. 67–8, argues strongly on this point.
48 See the discussion in part 5 of this chapter.

78 Sentencing aims, principles and policies
These points emerge from the sequel to the Birmingham mugging case of Storey
(1973).
49
Ayouth was ordered to be detained for 20 years for his part in the violent
robbery of a drunken man. The sentencewas widely publicized, both in Birmingham
and in the national newspapers, as an exemplary sentence. Researchers were able to
plot the rate of reported robberies in Birmingham and in two other cities during
the months before and after the sentence was passed. The robbery rates seemed
quite unaffected by the sentence in Storey: indeed, the rate of reported robberies
in Birmingham had begun to rise before the trial and continued to increase before
reaching a peak several weeks later. This calls into question the normal assumptions
one would make about human behaviour, unless it is argued that the effect of Storey
took several weeks to exert itself by reaching the ears of all potential robbers in
Birmingham. The difficulty is that we do not understand the reasons, and this
shows the problems of firm assertions about general deterrent effects.
The argument has returned to the empirical objection. The real test of the prin-
cipled objection is this: even if one believes the Notting Hill anecdote, would this
justify the extra-long sentences on the first people to be sentenced for the crime?
Should, for example, an extra two years of one person’s liberty be sacrificed in the
hope of deterring several others? The objection to this is often expressed in the
Kantian maxim, ‘a person should always be treated as an end in himself [or her-
self], and never only as a means’. Respect for the moral worth and autonomy of
the individual means that citizens should not be regarded merely as numbers, to be
aggregated in some calculation of overall social benefit. It may be true that the fun-
damental justification for the whole institution of punishment is in terms of overall
social benefit, in the same way as this is the justification for taxes. There are also
plenty of other examples of compulsion ‘for the greater good’, such as quarantine,
compulsory purchase of property and so on. These measures do not, however, have
the censuring dimension which sentences have. Exemplary sentences, by heaping an

undeserved portion of punishment on one offender in the hope of deterring others,
are objectionable in that they penalize an individual in order to achieve a social
goal – and do so without any real criterion of how much extra may be imposed.
Adeterrent theory which incorporates no restrictions to prevent this shows scant
respect for individuals’ choices and invests great power in the state and the judiciary.
There are several offences for which ‘deterrent’ rationales and sentence levels
are a long-standing feature – robbery and drug trafficking being prime examples.
The argument here is that it is necessary, in order to achieve a high level of general
prevention for such offences, to impose penalties which are more severe than the
proportionate sentencewould be. The Court of Appeal frequently upholds sentences
imposed on this ground; yet the empirical basis for expecting such policies to
succeed is almost entirely lacking – we do not know whether all courts impose such
sentences, whether this is known to offenders and potential offenders, whether
this knowledge affects their reasoning processes, or is outweighed by other reasons
49 (1973) 57 Cr App R 240.
3.3 The rationales of sentencing 79
(chance of avoiding detection, prospects of gaining substantial money). Moreover,
the judiciary seems confused on the issue. When Lord Taylor, as Lord Chief Justice,
was arguing against the introduction of mandatory sentences into English law, he
exposed the naivety of the government’s belief that such penalties would have a
significant deterrent effect, referring to the evidence against this and the evidence
that the risk of detection was more powerful.
50
Yetheand his successors as Lord
Chief Justice have presided in the Court of Appeal when many sentences based on
just such general deterrent reasoning have been upheld.
51
Governments are also
inconsistent on the point: only a few years earlier, a White Paper stated that ‘it
is unrealistic to construct sentencing arrangements on the assumption that most

offenders will weigh up the possibilities in advance and base their conduct on
rational calculation. Often they do not.’
52
Anumber of mixed theories of punishment have been advanced in an attempt to
preserve some elements of deterrence theory while avoiding the principled objec-
tions. The most notable is that of H. L. A. Hart,
53
who argued that the general
justifying aim of punishment must be found in the prevention and control of crime,
but that in deciding whom to punish and how much to punish the governing prin-
ciple should be desert. That is, only the guilty should be punished, and then only
in proportion to the seriousness of their offences. This does away with deterrence
as a rationale for particular sentences, but, on the other hand, it finds no place for
desert in the basic justification for punishment. There is a strong argument that in
order to justify punishment there must be insistence on individual desert as well as
overall social benefit.
54
Sentences are not the only form of general deterrent flowing from the criminal
justice system. In some cases it is the process that is the punishment – being prose-
cuted, appearing in court, receiving publicity in the local newspaper – rather than
the sentence itself. In some cases the shame and embarrassment in relation to family
and friends are said to have a more powerful effect than the sentence itself.
55
On the
other hand, the deterrent effects of sentencing and of the process may be diluted
considerably by enforcement policy, or at least by beliefs about the risk of detection.
As we noted earlier, the evidence suggests that it is beliefs about the probability
of detection rather than about the quantum of punishment which are more likely
to influence human behaviour.
56

However, there is little detailed knowledge of the
beliefs and thought processes of offenders and potential offenders, and the Cam-
bridge study indicates a need for more focused research on these matters.
57
Atatime
when the detection rate for all crimes has fallen to around 23 per cent, and when
burglary and robbery have detection rates of barely one-fifth, there are grounds for
believing that any deterrent effect which sentence levels have upon the reasoning of
potential offenders may be diluted considerably by the fairly low risk of detection. At
50 Taylor (1996), p. 10. 51 See n. 170 below. 52 Home Office (1990), para. 2.8.
53 Hart (1968). 54 Lacey (1988), pp. 46–56; von Hirsch (1993), ch. 2.
55 See the survey of young people by Willcock and Stokes (1963).
56 See the review by von Hirsch et al. (1999), ch. 6. 57 Von Hirsch et al. (1999), ch. 6.
80 Sentencing aims, principles and policies
any event, there is much less research in support of marginal deterrence by increas-
ing the severity of penalties: few such effects have been reliably identified, and there
are awkward questions such as how great an increase in severity is required, how
that can be communicated to the target audience, whether the severity of penalties
has already reached saturation point.
58
Thus, all the indications are that it is na
¨
ıve
to assume the kind of hydraulic relationship between court sentences and criminal
behaviour that some find intuitively appealing.
3.3.3 Incapacitation
59
Asecond possible rationale for sentencing is to incapacitate offenders, that is, to
deal with them in such a way as to make them incapable of offending for substantial
periods of time. In its popular form of ‘public protection’, this may be advanced as a

general sentencing purpose.
60
However, it is usually confined to particular groups,
such as ‘dangerous’ offenders, career criminals or other persistent offenders. Capital
punishment and the severing of limbs could be included as incapacitative punish-
ments, but there are formidable humanitarian arguments against such irreversible
measures. The debate has usually concerned lengthy periods of imprisonment and
of disqualification (e.g. from driving, from working with children, from being a
company director). Some community measures, such as curfews, may raise similar
problems.
What has been claimed for incapacitative sentencing strategies? This question
receives detailed discussion below in the context of persistent and ‘dangerous’
offenders,
61
but two such strategies can be mentioned here. One is the imposition
of long, incapacitative custodial sentences on offenders deemed to be ‘dangerous’.
It is claimed that one can identify certain offenders as ‘dangerous’, that is, as likely
to commit serious offences if released into the community in the near future, and
the risks to victims are so great that it is justifiable to detain such offenders for
longer periods. The chief objection to this is over-prediction: studies suggest that
incapacitative sentencing draws into its net more ‘non-dangerous’ than ‘dangerous’
offenders, with a ‘false positive’ rate that has often reached two out of every three.
This means that any portion of punishment added to the proportionate sentence
may be not only undeserved but also unnecessary to prevent that individual from
committing a further serious offence.
The empirical basis of the second incapacitative strategy is likewise open to ques-
tion. It was claimed by Greenwood in the United States that one can identify certain
high-risk robbers and incarcerate them for substantial periods, achieving a reduc-
tion in the number of robberies and lowering sentence levels for other robbers.
62

The
crime preventive benefits of this are obvious, but the strategy has been shown to have
major flaws. A subsequent report in the United States for the National Academy of
58 Von Hirsch et al. (1999), ch. 10.
59 For fuller discussion and selected readings, see von Hirsch and Ashworth (1998), ch. 3.
60 As in s. 142(1)(d) of the Criminal Justice Act 2003.
61 See ch. 6.7 and 6.8 below. 62 Greenwood (1982).