Crime reduction is high on the public policy agenda, particularly here in the UK against a backdrop of large funding cuts across Government. The UK Home Office has responsibility for reducing crime and criminal activity. So amongst other things, we’re talking about more police on the streets, more resources devoted to stopping things like cyber-crime and other white-collar crime, right?
Well, yes. That is right. Dead right, in fact. But there’s more to it than that. The Department for Education will never introduce policies with the explicit aim of reducing crime, but education policies can nevertheless contribute to reducing crime in the UK. There’s a small, but growing, body of literature showing that the relationship between education and crime in the UK is a causal one, and the benefits – in terms of the costs of crime avoided – from education policies that help to reduce crime can far exceed their costs, banking a net benefit for society.
In theory, education can affect crime through three main mechanisms:
- Income effects: education increases the payoff to legitimate work, which, in turn, makes working more worthwhile than criminal activity. Although it can also work in the other direction, as education can also increase the earnings from certain crimes (e.g. white-collar crime such as fraud).
- Short-sightedness: young people who leave education earlier tend to care more about today than they do for tomorrow – in other words, they prioritise short-term gratification in favour of long-term benefits (they have a high discount rate, in the jargon). This makes them more likely to undertake risky activities, such as crime. Better educated people tend to attach a higher value to their future, which means they’re less likely to engage in risky activities that could put their future in jeopardy.
- Self-incapacitation: time spent in school means less time on the streets committing crime. However, keeping young people in school who would have been likely to commit crime may also make them likely to commit violent offences whilst in school.
Is it possible to measure the effect of education on crime?
Yes, but there are data issues. For example, one of the most common measures of crime is the number of convictions, but this will not capture unsolved crimes or unreported crimes.
Another issue is that good data on crime and educational outcomes do not exist in the same place and so will have to be matched together, requiring suitable matching variables in each data source.
Finally, identifying a true casual effect of education on crime may be difficult. For example, those people who are likely to stay in school and get a good education may also have characteristics that make them less likely to go out and commit crime anyway. Taking account of these characteristics can be challenging, as they don’t often appear in datasets – i.e. they are unobservable. Unless everything that influences someone’s propensity to commit crime can be taken into account, one can never be sure if it’s the education improvement that’s driving any reduction in crime or something else that isn’t being observed.
Has anyone attempted to identify a true causal effect of education on crime? How big is it?
Due to the data issues mentioned above, there are few UK studies that look at the direct link between education and crime. There are just two studies of note, both authored by Steve Machin, Olivier Marie and Suciča Vujić.
The first, published in the Economic Journal in 2011, links crime data from the Offenders Index Database (OID) to education data from the General Household Survey. Machin and colleagues then attempt to elicit the causal effect of education on crime by using the raising of the school leaving age (RoSLA) in the 1970s to try and identify if a causal relationship exists.
They are able to do this because the changes in the school leaving age generated sharp differences in educational attainment among people born just months apart. Those pupils whose 15th birthday fell before the 1st September 1972 were allowed leave school at age 15. Those whose birthday fell on or after the 1st September had to stay on to complete an additional year of education. In effect, this is akin to a natural experiment in science: there’s no reason to expect the propensity to commit crime of the two cohorts of children to be different overall, not when they are born just months apart. So all else being equal, any reduction in crime following the RoSLA is likely to have a causal relationship with the extra education attained in that year.
Their results show that a 10% increase in the age at which people leave school would lower the number of convictions (per 1000 of the population) by 2.1%. They also show that reducing the proportion of people leaving school with no qualifications by 1% would reduce convictions for property crime by between 0.85%-1%. Their results concerning violent crime were largely statistically insignificant, which they say is likely to be because violent crime tends to be motivated by emotional rather than economic reasons.
Having identified that education has a significant and causal effect on convictions for property crime, Machin and his colleagues attempt to calculate the net social benefits of reducing the proportion of people in the population with no qualifications by 1%.
Although their cost-benefit calculation should be carefully interpreted, it is nonetheless insightful: a 1% reduction in the proportion of people with no qualifications could yield net economic benefits of between £23 and £30 million, ten years after the policy was introduced.
The second study by the trio, published last year, focusses on the link between youth crime and education (young people aged 16-21). The introduction of GCSE qualifications in 1988 (the qualifications UK schoolchildren get at the end of compulsory education at age 16) led to a significant expansion in post-16 full-time participation in education. Machin et al use this exogenous variation in post-compulsory participation to identify the causal impact of education on youth crime, using the same principles that apply to the RoSLA. They use matched data from the IOD (crime) and the UK Labour Force Survey (education).
For young men, they find a strong crime-reducing effect of education: a 1% increase in the proportion of male students reduces male conviction rates by around 1.9% and a 1% increase in the proportion of men staying on at school in post-compulsory education reduces male conviction rates by around 1.7%. The effects for women were smaller, at 1.1% and 1.3% respectively.
Interestingly, in this study the authors do find a positive crime reducing effect of education on both property crime and violent crime. They don’t particularly go into the reasons why this might be, but perhaps it’s because young people are short-sighted, (mechanism 2, above) and are driven more by their emotional responses at the time, rather than any economic reasons.
The authors also find that the expansion in participation positively affected wages and qualification attainment, which they interpret as meaning that the incapacitation effect is not the sole driver of the results – the income effect is at play here, too.
So, what can we conclude from this?
The first thing that jumps out is that this last finding doesn’t feel particularly satisfactory. Although the authors are probably right to conclude that the income effect is at play, they leave the reader wanting to know a bit more about the relationship between the two – which has the biggest effect on crime? In fact, all three mechanisms listed at the beginning of this article could be apparent, because staying on in education may also change the way that young people think about their future. Although Machin et al have no way to measure this in their data.
Identifying the relative strength of each effect is important; and from a policy perspective, the answer matters. If the income effect is small (and hence the results are driven principally by the other two mechanisms), then to an extent, it doesn’t matter what kids study, so long as they stay in school and come away with something. In turn, this implies that even low-level qualifications that have little labour market value can confer some wider benefits to society (since studying for / gaining them can help to reduce crime and the costs associated with crime).
The first study provides some support for this. In the UK, people who go from having no qualifications to something tend to start off with low-level, basic qualifications, with little labour market value. If Machin et al are right, helping people who have no qualifications to get at least something, could yield sizeable economic benefits for society.
Of course, this relationship is a complicated one. There could be many other factors at play here, not least the fact that low level qualifications could be used as a stepping-stone to something higher, which has more value in the labour market. At this point, the income effect becomes stronger.
Whichever side of the fence you come down on, the existence of a causal link between education and crime does strengthen the case for government intervention in the education market – particularly if low-level qualifications confer significant net benefits to society in terms of the costs of crime avoided.
To that end, building on the work of Machin et al and attaching a sensible value to this causal link between education and crime is important, because it helps policymakers to properly appraise and evaluate the impact of what they do. Without it, there’s a risk that those in Government could take too narrow a view of the costs and benefits of their policies. That’s bad for all of us, because it means that some policies that could confer significant wider benefits to society might not go ahead.
Machin, S. Marie, O. and Vujić, S. (2011): The Crime Reducing Effect of Education, The Economic Journal, 121 (May) 463-484.
Machin, S. Marie, O. and Vujić, S. (2012): Youth Crime and Education Expansion, German Economic Review.