Time to Educate the Criminals?
Kirstine Hansen

Introduction Figures published in Britain in July 2000 revealed an alarming rise in crime rates, renewing public concerns about the measures used to combat criminal activity. In the wake of the 2001 general election, the Blair government finds itself struggling to deal with a crime wave. In this article, Kirstine Hansen of the Centre for Economic Performance based at the London School of Economics and Political Science, reports on the role of education in tackling crime.

The figures make grim reading. In the year to March 2000, crime in the UK rose by 3.8 percent--5.3 million offences committed compared with 5.1 million offences in the 12 months to March 1999. What's more, authoritative sources suggest there may be widespread under-reporting of crime--that the actual figures could be as much as 20 percent higher. The rise brought to an end an unprecedented period of falling crime in the years after 1992.

Yet the new figures are consistent with the long-term trend in Britain (and, indeed, in most countries). Crime in England and Wales, for instance, has on average increased at about 5 percent a year since 1900; and grew even more rapidly in the 1970s and 1980s. This long-term trend has been remarkably impervious to changes in the criminal justice system, prompting social scientists to explore the underlying causes of crime and to look for fresh ways to tackle it. Attention has begun to focus in particular on the significance of crime committed by young people.

What do we know?

Leaving aside concerns about the possible under-reporting of crime, we actually know quite a lot about the crimes committed in England and Wales. In the year to March 2000, 83 percent of all crimes involved property: 50 percent of these were thefts, 20 percent were thefts of and from vehicles; and 8 percent were burglaries. Thirteen percent of all crimes were violent crimes; the remaining 4 percent included drug offences, public-order offences and crimes involving the prevention of the course of justice.

But the crime statistics only tell us what crimes were reported to the police: crucially, they do not tell us who committed the crimes. For that information we must look to the figures relating to those found guilty of or cautioned for indictable offences. (Since not all crimes are solved, these figures only relate to a proportion of all crimes reported.) In 1997, there were 509,000 offenders: the vast majority--82 percent--were male and about 25 percent were under 25. Figure 1 clearly shows that involvement in crime tends to rise and peak in the mid to late teens and early twenties. It is the rising level of youth crime, taking place at a time when the proportion of young people in the population is declining, which has become the focus of much public concern and, in turn, the efforts of policymakers.

Fig.1: The crime age profile of males aged 10-71, 1997.

Why the young turn to crime

In trying to combat the problem of youth crime, we first need to establish why young people are more likely to commit crimes. Whilst they are young most individuals have no strong sense of self-identity; much of their behaviour is based on trying to achieve short-term desires. Delinquency could simply be a way of getting 'kicks', having a laugh or relieving boredom. Peer pressure may increase delinquency as youngsters are encouraged to prove themselves and show loyalty to their peers. At this stage most youngsters feel little pressure to conform to societal norms, which means social controls are unable to deter them from breaking the law. What's more, young people tend to be protected from harsh punishment in the criminal justice system. This combination of factors would seem to be a recipe for high levels of youth delinquency and criminal activity.

But as young people grow older, they begin to be influenced by a series of factors which discourage them from breaking the law. They start to think of much delinquent behaviour as childish. As they move from dependence to independence, leaving school and the parental home and entering the labour market, getting married and starting families of their own, young people begin to develop ties to society and attachments to social institutions such as the family, the labour market and the community. These factors, coupled with the possibility of more severe legal sanctions, all encourage a lower crime rate, at least in public, as young people move towards adulthood.

Why aren't all young people criminals?

All this explains why on average young people are more attracted to crime than older people. It doesn't explain why two people of the same age don't display the same propensity to become criminals--if they did, then by definition all young people would turn to crime. This has led social scientists to explore what other factors, besides age, are important. One key factor (though not the only one) is of course exposure to the education system--a child's experience of school. New findings suggest this may play an important role in determining the likelihood of a young person's involvement in crime.

A 1998 report by HM Chief Inspectorate of Prisons showed that there were 10,570 young people under the age of 22 in the custody of the prison services in England and Wales in 1997. This represented a 5 percent increase on the previous year. The report points out that 'most of the youngsters had been failed by the education system'. Around two thirds of these youths had no formal qualifications, many had regularly played truant from school and over 50 percent had been excluded (or left voluntarily) before the age of 16.

These findings reinforce an important link between education and offending which has been found in many empirical studies. Take the importance of staying on at school, until the proper leaving age. An American study in 1999 found that high-school graduation reduced criminal participation among young males in the US, even after differences in ability were controlled for. It also found that young male high-school graduates were 30 percent less likely to earn an income from crime than those who did not graduate. Moreover, high-school graduation reduced the probability of being arrested by around 60 percent and of incarceration by between 85-95 percent.

The links between education and crime

Education can affect the likelihood of offending in a variety of ways. The cynical explanation is that whilst youngsters are at school they are being kept off the streets. This separates them from their most delinquent peers (who are likely to be absent from school) and enforces some level of discipline upon them. But there is more to it than that. Children at school are encouraged by the idea of meritocracy to have aspirations, to create goals which by working hard at school they will be able to achieve. This encourages children to develop a stake in their own future and in society more generally. All these factors would tend to reduce the involvement of young people in crime.

Perhaps most importantly, though, education encourages children to develop skills and acquire knowledge and training which will affect their future success in life. Their ability to communicate and forge relationships, the choices they make at the end of compulsory education, the jobs they will do and the wages they will receive over their lifetime potentially depend on the skills they acquire whilst still at school. If children want to maximise their future success they will be less likely to offend as youngsters. And if they secure successful jobs with good wages as a result of their educational success, they will also be less likely to offend as adults.

The crime-age profile

New research at the Centre for Economic Performance has underlined the importance played by education. We used self-reported data collected from young men aged 16-25 in England and Wales to examine the crime-age profiles of two groups: those who leave school at 16 and those who stay on past the compulsory school-leaving age. We found that the two groups have significantly different crime-age profiles: but that the gap between the two profiles can be accounted for by clear differences across the two groups in a number of observable variables related to the labour market, education, family, individual and the area/neighbourhood in which the young men live. Of these, the three most important are whether an individual lives with their parents, family contact with the police and school truancy. These tell us a great deal about why an individual stays on at school or not, and the consequent likelihood of their involvement in crime.

If the age at which an individual leaves school has no link with their involvement in criminal activity then we would expect the crime-age profiles of the two groups to be essentially the same. Figure 2 shows clearly that this is not so. For those who stay on at school, criminal activity is almost non-existent by the age of 25; for those who left school at 16, there is no real decline in the crime rate from the age of 22 onwards.


Fig. 2: Crime-age profile by education. Basic model for property crime.

On the face of it, these findings have important policy implications: could obliging people to stay on at school affect the crime rate? Unfortunately, the issue isn't quite so clear-cut. We need to know why these crime-age profiles are so different in order to determine the correct policy response.

In order to do this, we have examined other variables which might influence the crime-age distribution for the two groups and controlled for them to see what impact they have on the crime-age profiles. If any of these variables were able to account for a significant proportion of the difference in the crime-age distributions between the two groups, the two distributions would become more similar. If any or all could completely explain the difference in the profiles then the gap would be eliminated and the two groups would have the same crime-age profiles.

Where people live

Crime and delinquency are unevenly distributed. Evidence from the British Crime Survey suggests that over half of all property crime and a third of all victims of property crimes are found in just a fifth of communities in England and Wales. In the 1990s those in the worst crime areas suffered twice as much property crime as anyone else in England and Wales. The police statistics reveal similar trends. Police force areas that include large urban conurbations have the highest rates of recorded crime. In the year to March 2000, metropolitan forces recorded an annual rise of 7.2 percent compared to the 0.9 percent recorded by non-metropolitan forces. The greatest increases were in the West Midlands, which saw a rise of 16 percent, and in the Metropolitan Police area, with an increase of nearly 13 percent. These forces, together with the City of London, Greater Manchester and Merseyside, recorded increases totalling 190,000.

Within these broad areas, crime rates are highest in inner city areas, those with a high proportion of social housing, and poorly maintained districts. These patterns have been accentuated by recent trends in crime prevention and control which have encouraged individual self-protection, home security devices, neighbourhood-watch schemes, insurance and private policing which lead to increased crime in poorer areas, where individuals cannot afford to protect themselves.

It is not difficult to see that schools influence children's behaviour outside as well as inside the school, particularly in relation to delinquency. The age an individual leaves school, their attendance, whether they have been excluded and the qualifications they gain are all differentially associated with offending.

And who they are

Empirical work has shown that a number of individual characteristics are associated with offending. For example, crime rates are higher for non-whites than for white people. According to the Home Office, 18 percent of the prison population of England and Wales in 1997 was non-white men, even though non-whites accounted for only 6 percent of the total population of England and Wales. Marriage has also been found to discourage involvement in crime. And young people who have good relationships with their parents are less likely to be involved in criminal activity, while those who have run away from home are more likely to be offenders--as are youths living away from home.

Delinquents disproportionately come from lower-class and low-income families. Their parents, if in employment, tend to be in low-paid manual jobs. Delinquents are also more likely to have convicted parents or delinquent older siblings.

Several studies have examined the link between the labour market and crime. Although there is as yet no consensus, many found that, at least to some extent, crime is related to unemployment, inequality and low wages.

Which factors matter most?

When all these factors are taken into account, the crime-age profiles are indeed substantially altered. Figure 3 shows that the two profiles now look very similar. They completely come together at ages 16, 24 and 25; with only a slight gap between the two profiles from 17 to 23. All the variables combined account for 90 percent of the overall gap.


Fig. 3: Crime-age profiles with all variables: property crime.

Of course some of these characteristics matter more than others. Table 1 shows that the most important set of variables explaining the gap on average are school variables which account for approximately 46 percent of the gap overall; family variables which explain 26 percent of the gap; and individual variables at 27 percent. But we can also see from Table 1 that there are variations across ages. For example, individual variables explain very little of the gap at the younger end of the age profile (4 percent of the gap at age 16), but much of the gap between the ages of 20 and 22. This is perhaps linked to the movement away from the parental home, towards setting up new families and having children in the early twenties.

Being more specific

So we know that school variables, family and individual variables account for the largest proportion of the gap between the profiles for those with more education and those with less. But we would also like to know which specific factors within these groups are most important. And we do. At school, truancy is the most important factor. Those who have played truant are 14 percentage points more likely to commit property crimes. Within the family, what's important is whether an individual's family has had contact with the police: those with family members who have had contact with the police are 19 percentage points more likely to commit offences. And of the individual characteristics, what matters most is whether an individual lives with their parents at age 16. Those who do are 8 percentage points less likely to commit crimes.

The implications for policy

There's little doubt that these findings have important implications for long-term efforts to reduce crime committed by young people. One obvious solution might be to encourage more young people to stay on at school--and thus hope that young people can 'jump' from one crime-age profile to the other. In fact, we do not yet know enough to know if this would work, since we don't know what role is played by educational qualifications which those staying on acquire. The other route for policymakers to explore, of course, is in constructing social and educational policies which can have an impact on the specific factors which close the gap between the two crime-age profiles.

Neither approach offers a quick fix to the problem of youth crime. But set against a century-long trend of rising crime, the effort needed to come up with policies which can have a long-term impact is well worth making.

Relevant Link
Centrepiece Magazine


This feature has been adapted from chapter 7 in the book Children and their changing media environment, edited by Sonia Livingstone and Moira Bovill, Lawrence Erlbaum Associates, 2001.