Less crime without more punishment (*)
D. J. Smith (#)
Reproduced with permission from the Edinburgh Law Review (1999)
Vol. 3, pp 294-316
Return to Academic Papers
At its best, criminology is a
dialogue between theories of crime and evidence about behaviour and social
practices. The purpose of the
enterprise is to inform moral and political choices. A particularly important decision,
which ought to be illuminated by criminological theory and evidence, is
whether to punish more, or to use alternative methods of tackling crime.
This article argues that it is possible to have less crime without more
punishment. Comparisons between trends in
crime and punishment in the United States and England & Wales between 1981
and 1994 show an association between changes in the crime rate and changes
in the probability of punishment, given an offence. However, three lines of argument
can be developed against the proposition that changes in the probability of
punishment caused changes in the crime rate. First, it is more likely that the
causal arrow points the other way.
Second, detailed causal mechanisms have not been adequately
described, and there is little evidence to support any particular
description. Third, trends in
crime and punishment in Scotland do not fit well with the hypothesis. Crime survey results suggest there
has been little increase in crime in Scotland since 1981, without any
appreciable increase in the quantity or probability of punishment. The enormous post-war growth in
crime in the developed world, and the subsequent reversal of this trend, are
linked to a range of fundamental social and economic changes, which have
probably been far more influential than the criminal justice system.
A. WHAT IS THE POINT OF
In his inaugural lecture as Professor of Penology at the University of Edinburgh, David Garland gave an account of the essential nature of penology, its field of coverage, and its relationship to criminology. In doing so, he described penology as “the master discipline” and criminology as “a subsidiary field of study”.[i]
In my view, it is not important how the doctrines or observances of criminology are defined, or how this body of scholars sets itself apart from others. Criminology is not a sacred flame. It has no select band of adepts who alone have knowledge of its mysteries. It is not a strict system of thought allied to a single-minded way of life, on the pattern of a monastic discipline. Unlike, say, psychiatrists or advocates, criminologists do not have a powerful position grounded in the function of controlling the insane or pleading for the accused. They certainly do not form an élite club. They may well find themselves in a subordinate position. Intellectually, they do not constitute a self-standing discipline, but (as David Garland pointed out) feed off ideas and methods originating elsewhere.
It follows that it would not be fruitful to ask what is the essence of criminology. For, at its best, criminology is a group of people who use a bag of tools to think about crime and its theoretical and practical implications. Because the tradition is eclectic, the tool bags are filled with the instruments of many trades: history, sociology, psychology, psychiatry, the philosophy of law, and so on. In the short history of the subject, every attempted explanation (for example, for the post-war rise in crime) has proved inadequate. One theory or perspective, one method of research, has quickly followed another. The priority is not to maintain theoretical or methodological purity, but to provide some improvement on saloon bar theorising about pressing issues such as the treatment of young offenders, the use of imprisonment, or the response to racist violence.
Although it would not be fruitful to ask what is the essence of the discipline, it is perfectly reasonable to ask what is the point of criminology. The core of the answer is that it must always be a mixture of the practical and the mystical. This points to an analogy with the habits and observances of the Society of Friends, or Quakers. Friends have always concerned themselves equally with direct, practical action, such as relieving poverty, or treating the wounded while refusing to take part in war; and with meditation as a means of developing a deeper mystical awareness. These two aspects of the life of the Society are not unconnected, since mystical awareness is thought to inform personal moral choices, but this is not through any single system of principles that can be stated, disputed, or fought over. In fact, the only absolute Quaker principle is that there can be no absolute principles. Criminology also brings together the practical and the mystical. It is practical because the subject is defined by a field of application: what do we do about crime and criminal justice institutions? It is mystical because the answers will ultimately depend not only on facts and analysis, but also on personal moral choices which must be freely made, and cannot be forced on us by any set of principles. So the point of criminology is to provide understanding of crime and criminal justice institutions so as to inform moral and political choices.
Whereas there is no single, tightly-organised, fully coherent, or pure body of theory in criminology, there is certainly a rich fund of theoretical ideas. These can be used to address questions such as: Why do we punish? Why has the prison population in the USA increased fourfold in 20 years? Why is it so hard to change the police? The best site in which to address such questions is a meeting point between law, jurisprudence, psychology, and social and political theory.
In 1977, Derrick McClintock was the first person to be appointed to the chair of criminology at Edinburgh, the second to be established in the UK, after the first at Cambridge. At that time there were separate departments of criminology and jurisprudence. Derrick McClintock played a leading role in bringing them together into a single centre, and that combination of social, philosophical, and empirical studies remains an enduring part of his legacy. That environment favours the development of both theoretically informed studies of crime and criminal justice institutions, and a sociology of law and jurisprudence that engages with the actually existing social world.
The tradition of criminology that I wish to develop is one that consists of a dialogue between moral and political choices, theories of society and the individual, and evidence. Theories are modified or abandoned in the light of evidence; the search for evidence, and its analysis, are illuminated by theories; and the knowledge that arises from the interplay of observation and theorising is used to inform moral and political choices, without of course dictating them. In that way, scientific inquiry stimulates moral reflection, and leads to political consequences. This may well seem like a statement of the obvious. It is hard to see the point of theories of society if they cannot be evaluated in the light of evidence; equally, empirical research that does not help to explain the social world seems sterile; and scientific inquiry into the social world that does not help us to change it seems useless. And yet increasingly in the social sciences we see a separation between theory, empirical research, and practice, to the extent that a theoretician who advocates large-scale surveys, or a survey technician who proposes a theory, is considered a pretentious fool. The attempt to bring these things together is fresh and original, as well as necessary.
I would like to illustrate this general approach to criminology by considering the relationship between the quantity of crime and the quantity of punishment, and how the scientific results are related to the moral and political issues involved. I will argue that the amount of punishment is only one, rather unimportant, influence on the amount of crime, among many other more important influences. That simple conclusion—that most of the causes of crime are beyond control by the criminal justice system—forms the starting point for the new programme of research known as The Edinburgh Study of Youth Transitions and Crime. I will try to explain how this research programme embodies the general approach to social science set out above, and how it feeds into the political agenda of achieving less crime without more punishment.
B. DOES THE QUANTITY OF PUNISHMENT DRIVE THE CRIME RATE?
Does the quantity of punishment drive the crime rate? If we punish more, will that reduce the level of crime? If we punish less, will that lead to a rise in crime? Here is a question that has far-reaching practical implications. If more punishment means less crime, then perhaps the state would be justified in pouring more resources into law enforcement as opposed to crime prevention. Building prisons should perhaps take precedence over pre-school education. The same question also highlights a potential moral conflict. By and large punishment harms the individuals who are punished and reduces their capacity to support themselves and help others. That was most obvious when punishment involved lopping off limbs, but it is still sufficiently obvious where it involves imprisonment or confiscation of property. If these things are to constitute punishment, they must be understood as expressions of formal censure, and they must be accompanied by stigmatisation: the condemned person must be marked out by most people as a wrong-doer who is to be shunned and rejected, at least for a period. Of course, crime harms people too, but what is proposed is a Mephistophelian bargain, whereby we agree to do evil to the condemned in return for a reduction in risk for everyone.
But will the devil keep his side of
the bargain? Let me first
approach the question by considering historic crime trends.
(1) Explaining crime trends
From 1800, police-recorded crime in England rose consistently up to about 1855, and this may have been the continuation of a much longer trend of increase. Around 1855, there was a major reversal, and from that date, police-recorded crime fell consistently and substantially during the rest of the nineteenth century.[ii] By 1900, the decline had become gentle, but it continued until around 1935, when the next major reversal occurred. Crime then increased from 1935, gradually at first. Immediately after the Second World War there was a short-lived spurt in recorded crime, which led to political talk about a ‘war on crime’ now we had won the war against the Germans.[iii] It was not until the mid 1950s that the rise in crime began to accelerate. This became by far the most dramatic increase in crime since records began. Notifiable crimes recorded by the police (which do not include minor assaults, public order, and motoring offences) increased tenfold over the period from 1950 to 1993. Crimes of violence against the person increased by 33 times. Robbery increased by 60 times.[iv] In 1993 came the third major reversal over the period of 200 years. Quite suddenly, notifiable offences began to decline. Violent offences, which are a fairly small proportion of the total, continued to rise for longer. They may now be declining too, although it is too early to be sure.[v]
In the 1960s and 1970s, criminologists argued that the increases in recorded crime indicated a growth in the resources and attention devoted to controlling crime rather than an increase in crime itself.[vi] That position has now been abandoned by all serious scholars, for a number of reasons. Since 1973 in the US and since 1981 in Britain, the crime survey has provided an alternative measure of the level of crime. Only about one-quarter of the incidents revealed by crime surveys are recorded by the police. Yet, the crime survey results have tended to confirm and validate the recorded crime statistics. They suggest that the official statistics exaggerated the post-war rise in crime, but confirm that very substantial increases occurred from the time when the crime surveys began. Crucially, both the crime survey and the recorded crime statistics show the major reversal in the mid-1990s, although the survey shows it coming a couple of years later than the official statistics.[vii]
In summary, the evidence for England shows that over the past 200 years there have been very marked downward and upward trends in crime at different times; but the most striking feature of this pattern is the dramatic rise in crime in the period from 1954 to 1994, and its eventual reversal. This post-war growth in crime was so great that it must be regarded as a fundamental social transformation. Looking across the developed world, we find a growth in recorded crime of the same order of magnitude in every country, with the single exception of Japan. It might be suggested that the contrast between Japan and the rest of the developed world arises because of system differences or cultural differences. Differences between legal systems make it hard to compare like with like. Policing in Japan makes far more use of family and communal controls than in the west, so that the formal process of law is probably more often bypassed. Organised crime in Japan is so well-organised, and so entrenched, that it perhaps leaves few official traces.[viii] These differences might explain a lower rate of recorded crime in Japan at one particular point in time, but they cannot explain the actual pattern observed. For in 1950, the rate of recorded crime was the same in Japan and England. Presumably, the cultural and system differences between Japan and England were at least as great in 1950 as in 1993, yet the levels of recorded crime in the two countries at that time were the same. But between 1950 and 1993, when the crime rate rose tenfold in England, it remained level in Japan, so that by 1993 recorded crime was ten times as high in England as in Japan.
A central question for criminology is
whether these large and far-reaching changes in the amount of crime can
be explained—in whole or in part—by changes in the level of formal
sanctions: in the number or proportion of offenders who are caught and
punished. Yet in order to tackle that
question, we need to set the study of crime trends in a wider context.
Between 1992 and 1995, Sir Michael
Rutter and I conducted a collaborative study of trends in psychosocial
disorders in young people, in collaboration with a dozen scholars from
countries in Europe and North America.[ix] Most crimes are committed by
adolescents and young adults, so that the peak age of offending comes at
around the age of 18 in every culture, and at every historical period.
In our study, Rutter and I tried looking at crime as just one of a
number of disorders that are characteristic of young people.
The other disorders that we studied were abuse of alcohol and
illegal drugs, depression, suicide and suicidal behaviour, and the
eating disorders anorexia and bulimia.
We found compelling evidence that across a wide range of countries
there had been sharp increases in all of these disorders among young
people during the post-war period (except that this evidence was
relatively weak in the case of anorexia and bulimia). There is also evidence of links
between some of these disorders: for example, between depression and
suicide; between alcohol, illicit drugs, and crime. If we place the growth of crime in
the context of a more general increase in youth problems, it immediately
becomes less plausible that the ineffectiveness of the criminal justice
system is a major part of the explanation for it.
Our study revealed yawning gaps in
the evidence needed to test competing explanations for this striking
increase in psychosocial problems among young people. The central problem is that over
the post-war period western societies have been simultaneously
transformed in a number of different, though related, ways. It is inherently difficult to test
the hypothesis that one specific societal transformation was among the
causes of the rise in crime, whereas another transformation that
happened at the same time was not.
Difficult, but not impossible in principle. In the present state of the
evidence, Rutter and I thought that a number of broad conclusions could
be drawn (I am applying these to the rise in crime, although they mostly
also apply to the other psychosocial disorders too).
1. Any satisfactory explanation must deal with the unprecedented growth of crime during the golden era of prosperity, growth, and low unemployment between 1950 and 1973. In many or most countries, economic growth at this time was associated with narrowing income differentials and a reduction of inequality in living standards due to the introduction of the welfare state. Thus, although inequality may somehow be relevant to the growth of crime in the later period, especially after 1979 in England, it cannot be the major explanation for the post-war growth in crime.[x]
2. Although actual increases in inequality or deprivation cannot explain the growth of crime, there may have been an increase in awareness of inequality, in feelings of injustice, and in people’s vision of unattainable possibilities. These feelings may, perhaps, be especially strong in young people. They may be associated with a growth in consumerism, education, travel, and the mass media.[xi]
3. The rise in crime is much more clearly associated with economic growth, with increasing wealth and consumption, than with poverty or inequality. The explanation is provided by the burgeoning tradition of research which shows that crime is a response to opportunity. Economic growth created a vast range of new criminal opportunities over the post-war period.[xii]
4. A huge tradition of research on child and adolescent development has demonstrated that the way that families work has a powerful influence on the way that children mature and develop into adults. Parents who understand and look after their children’s needs, who are warmly affectionate towards them, who try to make sure they know what their children are doing, who are consistent in what they will and will not allow, who do not use harsh, arbitrary, or especially physical discipline, who are involved with their children and do things with them, who are there when they are needed, who are not in continual conflict with each other, are much more likely to produce children who stay out of trouble than parents who do not behave in these ways.[xiii] Although this remains controversial, it seems very likely that the rapid and far-reaching changes in the family that have happened over the post-war period[xiv] have been associated with an increase in the proportion of families that function badly, and hence a decline in the quality of parenting. This could be a major cause of the rise in crime.
5. A substantial proportion of crime is associated, directly or indirectly, with alcohol and illegal drugs, so that the rise in crime partly follows from the increase in alcoholism and drug abuse among young people, although of course to some extent the two kinds of disorder have common causes.[xv]
6. In the 1920s and 1930s, the first Chicago School established that youth crime rates varied widely between different neighbourhoods in the city, and that these neighbourhood contrasts remained constant as each new wave of migrants succeeded the previous one. The contemporary Chicago School is reinterpreting these neighbourhood differences. A recent article by Robert Sampson and others has shown that residents in Chicago neighbourhoods with lower crime rates are more likely to believe that local people can intervene to control the way others behave on the streets and in the parks.[xvi] In the jargon, it is the capacity for informal social control in public and semi-public spaces which varies between high-crime and low-crime neighbourhoods. Applying this insight to the explanation of crime trends, it seems likely that the capacity for informal social control has tended to decline in most communities over the post-war period.
There is a seventh type of explanation which Rutter and I tended to
reject, but which I now think may be a powerful one. This springs from the thought that
teenagers are a new institution, and that being young is more difficult,
dangerous, and tantalising than in previous ages. Throughout the twentieth century
there have been changes in the length and structure of the transitions
negotiated in adolescence, and an associated increase in the
distinctiveness and cultural solidarity of young people as a group. Adolescence is a much longer
period than before, because puberty comes earlier, whereas full
participation in the economy comes later. A consequence is that young people
remain economically dependent on their parents for longer, while at the
same time this dependency conflicts with a growing emphasis on the value
of individuality, self-expression, and youth culture. For a considerable period of their
lives, young people now spend most of their time in the company of other
young people, whereas at earlier periods of history they were usually
with adults. This
segregation reduces supervision and increases crime opportunities. It nurtures the growth of
distinctive youth cultures, which are given further impetus by
advertisers who target specific groups of young people as niche markets. The growth of education, the mass
media, and travel ensure that young people, have far more information
than before. Instead of
being quickly fitted into a pre-ordained slot in society, young people
now have a lengthy period of exploration, preparation, and indecision,
which they spend in the company of other young people having equally
(2) Assessing the influence of
I have deliberately set out to give a
set of connected explanations for the post-war growth in crime that
makes no mention of the criminal justice system at all. Crime trends are probably the
expression of complex, wide-ranging, and profound societal
transformations: surely they cannot be simply a response to the formal
system of law enforcement.
Yet it would certainly be wrong to dismiss the criminal justice
system out of hand as a relevant factor in explaining changes in the
level of crime. That point
has been brought into sharp focus by a series of articles and reports by
David Farrington, Patrick Langan, and P.-O. Wikström.[xvii] The starting point for these
comparisons was the contrast in crime trends between the US and England
in recent years. The rapid
increase in recorded crime in the US began in the early fifties, some
three years earlier than in England.
As in England, the growth was explosive in the golden era of
economic prosperity and low unemployment between 1950 and 1973. However, in the US, the reversal
came much earlier, around 1980, although the date varies depending on
which offence is considered, and whether recorded crime or crime survey
statistics are preferred.
In fact, for most categories of crime, there has been no increase in the
US since 1980, and for many there has been a substantial decline.
At the same time, from the 1970s
onwards, the US began its imprisonment binge, which quadrupled its
prison population in 20 years, and produced a rate of imprisonment now
five times as high as that of Britain, seven times that of Germany,
eleven times that of Norway.[xviii] Indeed, for locking up its
citizens, America is rivalled only by Russia, so that the former great
powers of the Cold War are now the great powers of incarceration. The reversal of the upward crime
trend in the US coincided with the expansion of imprisonment and an
increase in the number of convictions. Against everything else that I
have said so far, this seems like prima facie evidence that law
enforcement was an important influence on the crime rate. Shall we be prepared to boldly go
where hardly any British social scientist has gone before, to think the
unthinkable: that Michael Howard was right? That prison works?
Jeremy Bentham, when he first set out
a theory of criminal deterrence, made a sharp distinction between the
probability and severity of punishment. His purpose was to argue against
arbitrary, rare, and barbaric punishments, and in favour of more
moderate and certain ones.
He believed, like many deterrence theorists after him, that a likely but
moderate punishment would have a greater deterrent effect than an
unlikely but severe one.
Farrington and Langan have followed Bentham by conducting separate
analyses of the probability of conviction, given an offence; the
probability of imprisonment, given an offence; and the average length of
prison sentences. They
interpreted the results as showing that the volume of crime was
associated with the probability of conviction, but not with the
proportion of offenders sent to prison, or with the length of prison
sentences. This seems to fit with Bentham’s
view that certainty is more important than severity. In the following analysis, only
the probability of conviction will be considered.
Figures 1-8 illustrate Langan and Farrington’s most recent findings for England and Wales and the United States. The bars show the crime rate according to the national crime surveys, as referenced on the left-hand axis. Survey crime rates are based on answers given by large, representative samples of the adult population. For the categories of crime considered here, they provide a far more complete count than the statistics of crimes recorded by the police, since many incidents are not reported to the police or, if reported, are not recorded by them. Crime rates are produced by relating the number of incidents to the individual or household population, as appropriate. The line in each graph shows an index of the probability that an offence will lead to a conviction (referenced on the right-hand axis). This index is simply convictions as a percentage of the survey-based estimate of crimes.[xix]
In Figure 1, the bars show a steep rise in the rate of burglary in England and Wales between 1981 and 1995. The line shows the steep fall in the probability that a burglar would be caught and convicted. The pattern suggests that because more burglars were getting away, the rate of burglary increased. Figure 2 shows the comparable data for the USA. The England and Wales pattern is inverted. As the rate of burglary plunged, so the probability that a burglar would be caught increased.
Figures 3-8 show further findings of
this kind for three more offence categories. In the case of assault, the
pattern in England and Wales is the same as for burglary, and again the
US pattern inverts the one in England and Wales. For motor vehicle theft, the
pattern in England and Wales is much the same as for burglary and
assault, although the US pattern is rather less clear than for the other
offences. The pattern for
robbery is less dramatic, but still consistent with the hypothesis that
probability of conviction is inversely related to the rate of
In their first article on this theme in 1992, Farrington and Langan reviewed a number of alternative explanations for the divergent crime trends over the 1980s in England and Wales and the United States, and found none of them convincing. Without coming to a definite conclusion themselves, they seemed to invite the reader to conclude that the increasing probability of conviction or incarceration in the US had caused crime to level off or fall, whereas the declining likelihood of sanctions in England had caused crime to continue rising.
(3) Two counter-arguments
At least three broad lines of counter-argument are available to those of us social scientists who wish to keep a clean sheet in our opposition to Michael Howard. The first draws attention to the order of causation. The Farrington-Langan studies describe an association between probability of punishment and crime rate: and every first-year undergraduate in social sciences is taught that an association is utterly different from a cause, even if his teachers routinely ignore that qualification in their own research. Indeed, an entertaining paper in The Annals of Improbable Research claims that there is a strong correlation, in the US, between the risk of tornadoes in a state and the number of caravan parks. It explains this correlation by the theory that tornadoes seek and destroy caravan parks.[xx] In the present case, the question is whether changes in the probability of punishment drive the crime rate, or changes in the crime rate drive changes in the probability of punishment. The example of Japan will help to illustrate the point. There the resources devoted to law enforcement—the number of police per head of population, for example—are broadly similar to those in European countries. As I mentioned earlier, the crime rate in Japan, however, is very much lower. Not surprisingly, the police are much more successful than in Europe in catching the very much smaller number of criminals. Clearly the low crime rate leads to a much higher probability of sanction than in Europe. Equally, the higher probability of sanction presumably helps to maintain the low crime rate. In the jargon, we have bi-directional causation. It then becomes extremely difficult to devise statistical models to estimate the strength of the effect in each direction. It is perfectly possible that in the US crime levelled off and fell for reasons unconnected with sanctions, and this fall in crime then led to an increase in the effectiveness of formal law enforcement: in that case, the main causal arrow points from the crime rate to sanctions, not from sanctions to the crime rate.
Another possibility is that there is a statistical association but no causal relationship in either direction. In the US, crime rates may have started to fall for reasons unconnected with sanctions, while a political movement emerging in response to earlier rising crime rates eventually gave birth to increasingly punitive policies when the tide had already begun to turn. To take a specific example, ‘three strikes and you’re out’ was introduced after drug use had peaked.
The second line of counter-argument draws attention to the mechanisms through which sanctions are supposed to influence the crime rate. A statistical association is not enough to establish a causal explanation, because there has to be an intelligible mechanism connecting cause and effect, a mechanism that fits with what we know about psychology and social institutions. Deterrence theorists assume that rational choice has at least some relevance to the explanation of offending. On this model, potential offenders weigh up the costs and benefits of committing a crime, although this might be only one of a number of influences on them. Granted that there is an important element of rationality in decisions about whether or not to offend, it does not follow that marginal increases in the probability or severity of punishment will necessarily lead to reduced offending. One reason is that people have limited knowledge about the chances of being caught and what the punishment would be, and they place their own interpretations on the information they have. Another is that even if they recognise that the risks have increased, this may not be enough to put them off: it depends on how they subjectively evaluate the risk compared with the benefits to them of committing the crime.
Therefore, it needs to be demonstrated that potential offenders are influenced by an assessment of the risk of being caught and punished, and that marginal changes in the proportion of offenders who are punished, or in the severity of sanctions, influence people’s perceptions of the risks of offending. In fact, there is a considerable recent tradition of research on offenders’ decision making and their perception of risks, which has produced contradictory results depending on the research method used. There is evidence that things that offenders directly observe or experience affect the vulnerability of targets: for example, cars with central locking are less likely to be stolen, public telephones in positions where they are overlooked are less likely to be vandalised. But there is less evidence that changes in law enforcement activity or prosecution or sentencing policy have an effect on offenders’ perceptions of risk.[xxi]
Of course, the simple rational choice model overlooks some of the most basic points about the decision making process. Offenders’ decisions have to be seen in the context of the individual’s life history, or criminal career. The teenager who commits a few minor and casual offences is making one sort of decision. The young person who goes on to commit further offences after having been once caught makes a different sort of decision: the decision to become involved in a life of crime. The person with a long criminal history who goes out to commit a two hundredth offence makes an entirely different kind of decision. Finally, the criminal who decides after some years of offending to go straight makes another kind of decision again. Marginal changes in the likelihood or severity of sanctions probably have entirely different effects on these various kinds of decision.
The simple rational choice model also overlooks the vital relationship between the formal process of law, and informal social controls. Andrew von Hirsch has usefully developed the distinction between censure, shame, and stigma. Punishment is a communicative act that conveys censure, or moral disapproval, as well as directly causing pain or deprivation. Shame is the response of the person punished to being publicly censured. Stigma derives from the response of the general audience, who reject and shun the punished person as a wrong-doer.[xxii] Punishment therefore depends for its effect on the response of the individual and the audience, otherwise, as von Hirsch has observed, being sent to prison would be no worse than joining the crew of a submarine. Shame and stigma are strong as long as punishment is relatively rare. Increased use of the formal process of law dilutes the informal processes of stigmatisation that give meaning and force to punishment. In some US cities, a substantial proportion of the young black population has served a sentence of imprisonment. This must mean that among young black people in these neighbourhoods, little or no stigma is attached to being sent to prison. It follows that marginal increases in the amount of punishment have diminishing effects.[xxiii]
(3) The example of Scotland
These first two counter-arguments
focus on the direction of causation and the detailed mechanisms
involved, particularly perception of risk, and the shame and stigma
arising from punishment.
The third counter-argument revolves around finding counter-examples. Far too little use is made of
comparative research within the British Isles. It turns out that a crucial
example is available very close to home. The rise in recorded crime in
Scotland since 1950 was considerably less than in England. For example, between 1950 and
1995, recorded robberies increased by 15 times in Scotland, compared
with 67 times in England; serious assault increased by 17 times in
Scotland, compared with 39 times in England;[xxiv]
sexual assaults rose by 5 times in Scotland, whereas rapes increased by
20 times in England; and housebreaking increased by less than 3 times in
Scotland, whereas burglaries increased by over 13 times in England. In Scotland, the increases were
considerably greater over the first 21 years, from 1950 to 1971, than
over the following period from 1972, whereas in England the increases
were greater over the later period.[xxv] Crime survey results provide
further evidence that crime started to level off in Scotland some 15 to
20 years earlier than in England.
For all survey crime[xxvi]
there was no net change in the crime rate in Scotland between 1981 and
1995, whereas in England the rate increased by 60 per cent (see Figure
9). There was no net increase in
violence or acquisitive crime considered separately in Scotland over the
same period, although acquisitive crime appeared to rise and then fall
(see Figures 10 and 11).
At first sight, then, it looks as though Scotland may provide a crucial example that disproves the hypothesis that changes in the level of punishment drive the crime rate. As Peter Young and I pointed out in a recent article, there is no evidence that the law has been enforced more rigorously or punitively in Scotland than in England, and yet the crime rate has risen much more slowly in Scotland.[xxvii] There was a decline, rather than an increase, in the number of criminal charges brought in Scotland over the period from 1974 to 1994, when crime levelled off. The average daily prison population in Scotland fluctuated around the 5,000 mark between 1970 and 1990, and only then began to rise quite steeply. These broad statistical trends are in line with changes of policy. Scotland has developed policies to divert young offenders from prosecution, most importantly through the Children’s Hearing System. There was also a large increase in the diversion of adult offenders from prosecution before and during the period when crime levelled off in Scotland. At a time when the US was moving decisively and fast towards greater punitiveness and higher imprisonment rates, Scotland was moving firmly in the opposite direction. In both countries, rising crime levelled off.
More detailed comparisons of trends in Scotland, England and the US add a new twist to the argument. Figures 12-15 show for Scotland trends in survey-based crime rates and probability of conviction on the pattern of those shown in Figures 1-8 for England and Wales and the United States.
The rise in housebreaking between 1981 and 1987 was accompanied by a modest fall in the probability of conviction, but the fall between 1987 and 1995 was accompanied by no change in the probability of conviction (Figure 12). In the case of motor vehicle theft, Scotland showed a gentle decline in crime rate without any increase in the probability of conviction (Figure 13). In the case of assault, changes in both crime rate and probability were only slight (Figure 14). Robbery is the only one of these examples where the pattern of change in Scotland was consistent with the Farrington and Langan hypothesis (Figure 15).
These new findings for Scotland certainly lend no consistent support to the hypothesis that changes in the probability of punishment drive the crime rate. Contrary to my expectation, however, they do not clearly disprove it either. Of course, those of us who are implacably determined to harry the ghost of Michael Howard can go on looking for a more conclusive counter-example. Canada may be one, for there crime levelled off in the 1980s without any general increase in punitive law enforcement. However, I am inclined to think that we may be caught in a fork. If, as I have argued, a declining crime rate does lead to more effective law enforcement, then we will not be able to find an example where crime rates have declined without any increase in probability of conviction, but this will not mean that the level of punishment is driving the crime rate.
C. THE BALANCE BETWEEN LAW ENFORCEMENT AND PREVENTION
In intellectual terms, the debate I have been pursuing is about the primacy of formal law enforcement as an influence on the level of crime. In policy terms, it is about finding the appropriate balance between formal law enforcement and crime prevention. I do not mean to deny that a functioning criminal justice system, and a reasonably effective apparatus of law enforcement, are necessary conditions for most other efforts to reduce crime. The point at issue is not whether there has to be some law enforcement, and therefore some punishment: I assume that there must be. In Herbert Packer’s phrase, punishment is lamentable, but necessary.[xxviii] Instead, I have examined the claim that further increases in the amount of punishment have been the main reason for the end of the post-war rise in crime in the USA and Scotland. I believe I have shown that that claim is not well-founded. The counter argument is that a wide range of institutional and cultural changes influence the level of crime, and that these are jointly far more important than the level of formal law enforcement. There is a statistical association between the amount of crime and the likelihood that an offender will be punished, but this is primarily because rising crime makes it more difficult to catch offenders, whereas falling crime makes it easier. I have to admit that this whole argument would be immensely strengthened if we could identify causes outside the criminal justice system for the reversal of the post-war rise in crime in the USA and Scotland, and much more recently in England. As Peter Young has pointed out, explaining falling crime is the next major task for criminologists.
Meanwhile, as conscientious scientists, we do not have to believe that punishment is the main thing driving the crime rate. That leaves the way open for the further development of a Scottish model that increasingly emphasises alternatives to punishment: diversion from prosecution, particularly in the case of juveniles; early intervention to prevent the development of criminality in children and young people; situational crime prevention to reduce the specific opportunities for crime; community-based initiatives which aim to develop the capacity of communities to regulate their members. Of course, the scientific evidence does not force this choice on us. We are free to choose more punishment if we wish. But if more punishment were a medicine, doctors would not prescribe it, because it has not been shown to have consistent, general effects. At the same time, we know that its harms go well beyond the side-effects of useful medicines. We know that punishment is harmful because, paradoxically, it is only to the extent that the condemned person is harmed that the community benefits. We are free to choose instead to shift the balance to prevention. Criminological arguments and evidence can help to inform that choice, but do not dictate it.
D. THE EDINBURGH STUDY
The argument pursued in this article is a preamble to a programme of research, The Edinburgh Study of Youth Transitions and Crime, which started in January 1998. The programme is designed to build our understanding of a wide range of influences on criminal behaviour, and to lay the foundations for progressive and integrative policies of crime prevention. This programme, which is funded by the Economic and Social Research Council, has the ambitious aim of following the life histories of all children in the first year at Edinburgh secondary schools in 1998-99, a group of around 4,500. It is intended that the study will follow these children from the age of 12 to their mid-twenties at least. All state schools in Edinburgh, and most of the independent schools are taking part. This design, which covers a complete age cohort in a city, is unique among longitudinal studies focusing on crime. It has the advantage that all neighbourhoods, social strata, and social settings, are included. The study collects information once a year from the children themselves, from the schools, from the children’s hearing system, and from the social work department. As well as studying the individual children, the study will map the social geography and spatial crime patterns of the city, and aims to understand how children’s lives are influenced by the neighbourhoods where they live. A further aim is to study family influences through a survey of the parents of the cohort of children.
In a number of ways, this research programme embodies the conception of criminology set out here. It is designed to stimulate the dialogue between theoretical ideas and evidence. For example, it particularly aims to explore why male and female experiences of crime are so completely different. It is concerned with the details of individual experiences, perceptions, and decisions, and with how these fit with larger-scale theories, such as those about the effects of punishment. It seeks to link different levels of explanation: for example, to show how individuals are influenced by neighbourhoods, and how neighbourhood characteristics emerge out of relationships between individuals. Finally, it addresses the wider causes of crime, and therefore illuminates debate about prevention and intervention beyond the ambit of the criminal justice system.
I have set out evidence that over the past 15 years rising crime levelled off in Scotland without any marked increase in punishment. Scotland contrasts with England, where crime continued to rise strongly until 1993, and with the US, where the reversal of the upward trend was accompanied by a vast increase in imprisonment. At a minimum, these findings suggest that it is possible to have less crime without more punishment. How that can best be achieved is what we plan to investigate further. “Less crime without more punishment” is therefore a statement about the existing scientific evidence, and about the need for further evidence. It is also a slogan that advocates a particular moral and political choice. It is therefore both an intellectual and a policy agenda for the new Scotland.
+ An edited version of the inaugural lecture delivered on 4 May 1999 at the University of Edinburgh.
* Professor of Criminology, University of Edinburgh.
[i] D. Garland, “The punitive society? Penology,
criminology and the history of the present” (1997) 1.2 Edinburgh Law
[ii] For an account of nineteenth century crime
trends in England, see V. A. C. Gatrell, “The decline of theft and
violence in Victorian and Edwardian England” in V. A. C. Gatrell, B.
Lenman and G. Parker (eds), Crime and the law: The social history of
crime in Western Europe since 1500 (1980), 238-337.
Garland, “The punitive society”.
[iv] For an account of post-war crime trends, see D.
J. Smith, “Youth crime and conduct disorders: Trends, patterns and
causal explanations”, in M. Rutter and D. J. Smith (eds) Psychosocial
disorders in young people: Time trends and their causes (1995),
[v] For an account of recent crime trends in England &
Wales, see C. Mirrlees-Black, T. Budd, S. Partridge and P. Mayhew,
The 1998 British Crime Survey: England and Wales (1998).
[vi] E.g. I. Taylor, P. Walton and J. Young, The
new criminology: For a social theory of deviance (1973).
[vii] Mirrlees-Black et al, The British Crime
[viii] For a discussion of the low crime rate in
Japan, see D. Bayley, Forces of order: Policing modern Japan 2nd
edition (1991), ch. 9.
[ix] M. Rutter and D. J. Smith (eds) Psychosocial
disorders in young people: Time trends and their causes (1995).
[x] This was first clearly recognised by L. E. Cohen
and M. Felson, “Social change and crime rate trends: A routine activity
approach”, American Sociological Review 44 (1979), 588-608.
[xi] W. G. Runciman, Relative deprivation and
social justice: A study of attitudes to social inequality in twentieth
century England (1966). However, there are no adequate studies of
time trends in relative deprivation, so this suggested explanation for
rising crime is speculative.
[xii] R. V. Clarke, “Situational crime prevention”
in M. Tonry and D. P. Farrington (eds), Building a safer society:
Strategic approaches to crime prevention (Crime and Justice: A review of
research, vol 19) (1995) 91-150; M. Felson, Crime and everyday
[xiii] R. Loeber and M. Stouthamer-Loeber, “Family
factors as correlates and predictors of juvenile conduct problems and
delinquency”, in M. Tonry and N. Morris (eds), Crime and Justice: An
annual review of research vol 7 (1986), 29-150.
[xiv] L. E. Hess, “Changing family patterns in
western Europe: Opportunity and risk factors for adolescent
development”, in M. Rutter and D. J. Smith (eds), Psychosocial
disorders in young people: Time trends and their causes (1995),
[xv] J. Fagan, “Intoxication and aggression”, in M.
Tonry and J. Q. Wilson (eds), Drugs and crime: Crime and justice, a
review of research vol 13(1990), 241-320.
[xvi] R. J. Sampson, S. Raudenbusch and F. Earls,
“Neighborhoods and violent crime: A multilevel study of collective
efficacy”, Science 277, 918-924.
D. P. Farrington and P. A. Langan, “Changes in crime and punishment in
England and America in the 1980s”, Justice Quarterly 9.1 (1992),
5-46; D. P. Farrington and P.-O. Wikström, “Changes in crime and
punishment in England and Sweden in the 1980s”, Studies in crime
prevention, 2 (1993), 142-170; D. P. Farrington, P. A. Langan and
P.-O. Wikström, “Changes in crime and punishment in America, England and
Sweden between the 1980s and the 1990s”, Studies in crime and crime
prevention, 3 (1994), 104-131; P. A. Langan and D. P. Farrington,
Crime and justice in the United States and in England and Wales, 1981-96
[xviii] Criminal statistics England and Wales
1997, Cm 4162, at 213, Table 10B.
[xix] A considerable proportion of crimes are
committed by several people acting together. It follows that the actual
probability that an offence will lead to a conviction is different from
the ratio between convictions and offences. Langan and Farrington made
an estimate of the average number of offenders per offence (based on the
reports of victims in crime surveys) and combined that with the other
estimates to produce an estimate of the actual probability that an
offender would be convicted. However, estimates of the number of
offenders per offence are unreliable, especially for offences (such as
burglary) where there is usually no contact between victim and offender,
because they are based on a subset of incidents in which the victim saw
the offenders or had knowledge of who they were. Sample sizes on which
these estimates are based tend to be low, especially in the Scottish
Crime Survey. Because Langan and Farrington held constant the average
number of offenders per offence across the different years, it does not
influence the analysis of time trends. Therefore the simpler statistic
relating the number of convictions to the estimated number of offences
is preferred in the present analysis.
[xx] F. Wu, “The correlation between tornadoes and
trailer homes”, in M. Abrahams (ed), The best of annals of improbable
research (AIR) (1998), 82-84. Note that although some articles in
AIR are genuine science, this one is a spoof, even though it is making a
fundamentally serious point.
[xxi] For a review of the extensive wider literature
on criminal deterrence, see A. von Hirsch, A. E. Bottoms, E. Burney and
P.-O. Wikström, Criminal deterrence and sentence severity: An
analysis of recent research (1999).
[xxii] A. von Hirsch, Censure and sanctions
[xxiii] D. Nagin, “Criminal deterrence research at
the outset of the twenty-first century”, in M. Tonry (ed) Crime and
justice: A review of research, 23 (1998) 1-42.
[xxiv] The figures for Scotland refer to more serious or less
[xxv] These statistics are more fully documented,
with their sources, in D. J. Smith and P. Young, “Crime trends in
Scotland since 1950”, in P. Duff and N. Hutton (eds), Criminal
justice in Scotland (1999) 1-13.
[xxvi] Figures 9-11 actually relate to the
“comparable subset” of crime categories that can be closely compared
between the survey and the recorded crime statistics, but the following
statements would also apply to all incidents covered by the crime
Smith and Young, “Crime trends in Scotland”.
[xxviii] H. L. Packer, The limits of the criminal sanction (1968).