Education and the likelihood of crime
In the US, the net annual burden of crime is estimated to exceed $1 trillion. A large part of this sum is attributed to the cost of incarcerating more than 1.5 million individuals in state and federal prisons, and roughly 60,000 juvenile offenders in residential placement facilities (Carson 2014). The burden of incarcerating such a large part of the population extends beyond the financial costs of imprisonment, and includes decreased material and psychological wellbeing of the families of those imprisoned as well as lost productivity. In response to this significant social and economic issue, attention has increasingly turned to identifying the factors that lead to crime and incarceration. What has emerged as a consistent and robust predictor of these outcomes in the adult population is a low level of education. Recent research demonstrates that this relationship is causal, with higher levels of education reducing the likelihood of crime and incarceration amongst adults (Oreopoulos and Salvanes 2009). Juxtaposing this literature is a body of research that documents a trajectory leading to adult crime, in which the starting point is typically antisocial and delinquent behaviour in youth (Williams and Sickles 2002). Taken together, these two bodies of research suggest that the relationship between delinquency, education, and adult crime is a dynamic one, in which choices and outcomes regarding delinquency and schooling in youth are pivotal. Yet, the nature of the relationship between delinquency and schooling remains unclear.
Several recent studies have made useful contributions toward understanding this relationship. For example, arrests and incarceration in youth are found to reduce educational attainment (Kirk and Sampson 2013, Aizer and Doyle 2015), while school attendance decreases contemporaneous arrests, reported incidences of crime, and prosecutions for property crime (Berthelon and Kruger 2011, Anderson 2014). The criminal or juvenile justice focus of this literature has clear and important policy implications. However, interactions between youths and the justice system result from the decision of youths to engage in delinquency. This decision has the potential to impact on schooling outcomes, whether or not arrest or incarceration eventuates. For example, in a model of crime that accounts for human and criminal capital accumulation, experience in delinquency builds criminal capital stock which, in turn, lowers the returns to education relative to crime in adulthood (Lochner 2004). This makes education relatively less attractive, and an early school exit more likely, even if no arrest occurs. As a consequence, delinquency may lead to early school leaving for those who do not come to the attention of the criminal justice system, in addition to those who do.
Delinquency and early school leaving: New evidence
In recent work, we investigate the role of delinquency, in addition to arrest, in determining early school leaving using information on males from the National Longitudinal Survey of Youth 1997 (Ward et al. 2015). Figure 1 presents the transition (or hazard) rates into delinquency and arrest, and out of school (defined as primary, secondary or tertiary education). While the hazard rate for the transition into delinquency peaks at age 14, the hazard rate for first arrest peaks later, at age 18. In terms of school leaving, the hazard rate increases steeply from age 17 (shortly before most students graduate from high school at age 18 or 19), and peaks at age 19. It increases sharply again from age 21 (just before many students start graduating from college at age 22), peaking once more at age 24.
Figure 1. Transition rates for first delinquency, arrest and school-leaving by age
In summary, the hazard rate for the transition into delinquent behaviour peaks at an age younger than the peak in the hazard for arrest, which in turn occurs before the first peak in the hazard for school leaving.
Overall, Figure 1 shows that the initiation of delinquency and first arrest typically precede school leaving, and that should a causal relationship exist, it would likely run from delinquency and arrest to school leaving.
Establishing whether the effects of delinquency and arrest on school leaving are causal involves two challenges.
- First, unobserved characteristics such as time and risk preferences and cognitive and non-cognitive abilities that affect early school leaving are also likely to determine delinquency and arrest, rendering them endogenous to the school leaving decision.
For example, Heckman et al. (2006) show that higher levels of non-cognitive ability increase the probability that a person will graduate from high school and four-year college, while decreasing the probability of participating in crime.
- The second issue is that of reverse causality, with prior research showing that in addition to arrest leading to school leaving, school leaving increases the likelihood of arrest.
In order to address these issues, our research focuses on the transitions into delinquency and arrest, and the transition out of school using a multivariate mixed proportional hazard (MMPH) framework. Within this multivariate framework, we account for unobserved common confounders by introducing unobserved heterogeneity drawn from a joint distribution into the hazard rates for school leaving, the onset of delinquency, and first arrest. We account for reverse causality by modelling the relationships between the transitions into delinquency and out of school, and first arrest and school leaving as bi-directional.
Accounting for common unobservable confounders and reverse causality, we find that both arrest and delinquency increase the likelihood of early school leaving. Table 1 gives an indication of the magnitude of the effects for a reference individual (see footnote to the table). The first column presents the baseline case in which the reference individual has not engaged in delinquency, or been arrested. In the second and third columns the reference individual is assumed to initiate delinquency at age 16, with no arrest assumed in the second column and arrest at age 17 in the third column. The fourth and fifth columns present different scenarios. As shown, if the reference individual has not engaged in delinquency and not been arrested, the probability of leaving school at age 20 or earlier is 73%. This rises to 81% if he initiates delinquency at age 16 and is not arrested, and 92% if he has been delinquent at age 16 and arrested at age 17.
Table 1. Effect of delinquency and arrest on the cumulative probability of school-leaving for a reference individual who is susceptible to early school-leaving; simulations (%)
Note: Reference individual is a male who is susceptible to early school-leaving (along with 91% of our sample), is non-black and non-Hispanic, obtained a standardised (schooling-corrected) CAT-ASVAB score of 0.00, did not experience puberty prior to age 12, attended a public school, was born in 1982, did not have a teen mother at birth, does not have a responding parent (to the parent questionnaire) that is very religious, has a mother and father who are high school graduates (but have not attended college), has one younger and one older sibling, has a mother and father that are present, and resides in the suburbs, in the South.
In terms of the magnitude of estimated effects, being arrested for the first time is found to have roughly twice the effect of initiation into delinquency on the likelihood that an individual leaves school. However, amongst males who are enrolled in school at age 17, twice as many have been delinquent but not arrested compared to the number arrested. Therefore, the overall impact of delinquency on school leaving is at least as large as that of arrest in our sample.
More detailed analyses reveal that the effect of delinquency on school leaving is largely driven by initiation into delinquency that is income generating, and that early initiation into income generating delinquency has a larger effect on school leaving than later initiation. Similarly, first arrest leads to school leaving only if it occurs prior to age 18, and its effect is larger for younger ages at first arrest. These findings are consistent with predictions from a capital accumulation model of crime, suggesting that in addition to human capital, criminal capital accumulation may be an important mechanism through which delinquency impacts the school leaving decision. In contrast, we find that initiation into non-income generating delinquency only has an effect on school leaving if it occurs close to the end of high school, at age 18 or older. This may be evidence of salient or vulnerable ages with regard to the school leaving decision.
Our findings provide useful insights for policy development. Specifically, we show that there are a large group of delinquents who avoid arrest, but whose reduced level of educational attainment is as important as that of the group that have been arrested. As a result, to focus interventions solely on those who come to the attention of the criminal justice system would miss a large part of the vulnerable population. And while delinquents who avoid arrest may remain undetected by law enforcement, they are likely to have come to the attention of their school teachers and principals. This suggests that interventions targeting this group may be most effectively implemented within schools. For example, ‘Becoming a Man’ is a school-based prevention programme aimed at improving the social-cognitive skills of disadvantaged male youths in grades 7-10 from high-crime Chicago neighbourhoods. Further examples of such programmes include ‘Functional Family Therapy’, and the ‘PATHS (Promoting Alternative Thinking Strategies) Curriculum’. Given the huge costs associated with crime and incarceration in the US, there is much scope for effective school-based programmes to reduce these costs and improve the life outcomes of boys who are at risk of a future that involves the criminal justice system.
Aizer, A and J J Doyle (2015), “Juvenile incarceration, human capital, and future crime: Evidence from randomly assigned judges”, Quarterly Journal of Economics, 130(2):759 – 803.
Berthelon, M E and D I Kruger (2011), “Risky behavior among youth: Incapacitation effects of school on adolescent motherhood and crime in Chile”, Journal of Public Economics, 95(1-2):41– 53.
Anderson, D M (2014), “In school and out of trouble? The minimum dropout age and juvenile crime”, Review of Economics and Statistics, 96(2):318 – 331.
Carson, E A (2014), Education and correctional populations, US Department of Justice, Bureau of Justice Statistics, Washington, DC.
Heckman, J J, J Stixrud, and S Urzua (2006), “The effects of cognitive and noncognitive abilities on labour market outcomes and social behaviour”, Journal of Labour Economics, 24(3):411– 482.
Kirk, D S and R J Sampson (2013), “Juvenile arrest and collateral educational damage in the transition to adulthood”, Sociology of Education, 86(1):36 – 62.
Lochner, L and E Moretti (2004), “The effect of education on crime: Evidence from prison inmates, arrests, and self-reports”, The American Economic Review, 94(1):155 – 189.
Oreopoulos, P and K G Salvanes (2009), “How large are returns to schooling? Hint: money isn’t everything”, NBER Discussion Paper No. 15339.
Ward, S, J Williams, and J C van Ours (2015), “Bad behavior: Delinquency, arrest and early school leaving”, CEPR Discussion Paper No. 10755.
Williams, J and R C Sickles (2002), “An analysis of the crime as work model: Evidence from the 1958 Philadelphia birth cohort study”, Journal of Human Resources, 37(3):479 – 509.