Acton Institute Powerblog

Criminal justice reform: What does economics have to say?

Share this article:

This is part two of a series on criminal justice reform. Read part one here.

For many, crime and criminal justice are not obvious economic issues, despite their effects on public budgets due to the cost of courts, policing, investigating crimes, and much more. Private efforts impose significant costs, as well, whether from house alarms, flood lights, or door locks, not to mention the costs incurred by victims.

But costs such as these are not the primary source of economic interest in crime and criminal justice. If economics is a study of human behavior, the economic way of thinking can inform our understanding of the causes and effects of criminal activity and help us predict the consequences of public policy aimed at criminal justice. Therefore, economic literacy is key to contemplating the recent criminal justice reform debate and the First Step Act, which the House passed and President Trump signed into law just last week.

While the applicability of economics to crime is now readily accepted by today’s economists, this has only been the case for about 50 years. In a 1968 article, Nobel Laureate Gary Becker made a critical contribution to our understanding of crime, arguing that criminal activity could be modeled like so many other behaviors, based on the consideration of costs and benefits. Becker’s work was groundbreaking in that it provided a way for thinking about criminal activity and the people who commit crime within a generalizable model.

What provokes a person to commit a crime? The prevailing notion before Becker’s work was that crimes are committed by those of a criminal type. Becker proposed something different—that the higher the cost of crime or the lower the benefit of crime, the less likely an individual would be to commit a crime (and vice versa). That is, even those who commit crimes are likely to respond in predictable ways to changes in costs and benefits.

For example, a person who has meaningful employment faces a higher cost of incarceration in the form of foregone wages and fulfillment than does a similar individual without such employment. We would predict, then, that the former is less likely to commit a crime, other things held constant. Similarly, an individual who would have to separate from family if incarcerated would incur a greater psychological cost of crime, and be less likely than an individual with fewer interpersonal connections to commit a crime. As another example, the longer the probable sentence associated with a crime, the higher the cost to the potential offender and the less likely that person is to commit a crime. 

The length of criminal sentences has attracted a lot of attention as part of the debate about criminal justice reform. Some advocates assert that there are victimless crimes that are overcriminalized by the harshness of the sentences they receive. While many advocacy groups work for reducing or eliminating so-called mandatory minimum sentences, most are not entirely opposed to sentencing guidelines. Guidelines provide a predictability that some advocates consider consistent with justice, because a potential offender would know the likely consequences and understand that they should be applied impartially. The economically literate might also think that a potential offender’s understanding of the consequences of his criminal activity is important in the calculation of the net benefit of crime through the thought process modeled by Becker.

A Beckerian framework is not just helpful in predicting who might commit a crime or who will become a “criminal.” It is also helpful predicting recidivism risk for the already incarcerated, who will one day reenter society. Imagine a former offender who returns to society and finds a set of family, community, and employment attachments that are meaningful and rewarding. His likelihood of recidivism should be less than a similar person who finds himself with “nothing to lose” when faced with the prospect of another criminal act. This suggests that prison and sentencing reforms should be made with an eye toward re-entry. The provision in the First Step Act that would require placing a convicted offender in a prison within 500 miles of his primary residence would likely help to maintain family and community ties for incarcerated people. If an institution with appropriate security is available closer to an offender’s home, this is an appropriate investment in the offender’s wellbeing, his family’s wellbeing, and his later reentry prospects. In the final post of this series, I will detail more of the provisions of the First Step Act and the effects that economics would predict.

Building on Becker’s seminal work, a large literature within economics has identified causal connections between policy and criminal activity. With Dr. David Phillips of the University of Notre Dame, I coauthored an article on sentencing guidelines using a large dataset on two common crimes in Michigan. Our goal was to determine whether longer sentences always reduce future criminal activity. With data on retail fraud (including high-value shoplifting) and repeated drunk driving (third offense or more), we found that a more severe sentence did reduce recidivism for retail fraud offenders but not for those convicted of drunk driving. This result is consistent with a concern suggested in the first post in this series—that, where substance abuse or addictive behavior is involved, incarceration may not reduce recidivism as well as rehabilitation.

In general, economists have an important contribution to make to the discussion of criminal justice reform. First, economists are aware of the common social scientific adage that “correlation does not imply causation.” In the context of my coauthored research, it meant that we were not easily convinced that longer sentences were the cause of crimes later in life, even though people who received longer sentences did, on average, commit more crimes later in life. We suspect that judges see more serious crimes or more troubled offenders as indicating a higher risk of criminal activity in the future and respond rationally to that risk by lengthening the sentence. Economists are aware of this kind of challenge but also equipped to handle it with tools of statistical analysis less often employed by other disciplines or practitioners. 

Second, economists and the economically literate more broadly are aware that good intentions are not synonymous with achieving the desired results. Human behavior is both complex and rooted in self-interest, making it difficult to predict the full set of consequences for a policy, for example. The tough-on-crime era that preceded today’s popular “smart justice” movement was well-intended. It’s harsh sentencing regime also might have increased criminal activity through recidivism by reducing former offenders’ connections at home and in the labor market, creating the revolving door effect. 

Another example of an unintended consequence is found in many states that have made it illegal for employers to inquire about an applicant’s criminal history in the early stages of job application. The intention of these “ban the box” laws is to increase opportunity for former offenders by allowing them to get further along in an application process before disclosing their relevant history and, in some cases, reduce racial disparities in employment. As a forthcoming article by economists Jennifer Doleac and Benjamin Hansen finds, the result is far from what was intended. Specifically, African American and Hispanic men are less likely to be moved through the application process. The basic explanation is that, when employers wish to avoid interviewing former offenders but they cannot identify who has a criminal history, they will select for further consideration fewer applications from populations with higher average criminal history.

These lessons of economic literacy should inform the response to the First Step Act in its call for “evidence-based” programming in prisons. First, the prevalence of unintended consequences backs up the demand for a solid basis in evidence. Second, to tackle the causality challenge, evidence must be high-quality, based on large datasets, and discovered through the best statistical techniques available. It should also be rooted in solid economic theory that recognizes that incentives matter to people—whether former offender or not.

Image: Barbed Wire Chain Link

Sarah Estelle

Sarah M. Estelle, Ph.D., is Associate Professor of Economics in the Department of Economics and Business at Hope College in Holland, Michigan, and director of the Markets & Morality program. She is a research fellow at the Acton Institute.