The unprecedented surge in incarceration since 1980 has stimulated a national debate between those who claim that locking up over 2 million people is necessitated by public safety concerns, and those who say the human and financial burden of imprisoning so many of our citizens is intolerable.
Recent declines in some state prison populations do not reflect a “win” for prison-reduction advocates so much as the extraordinary stringency of state budgets resulting from the Great Recession. The issue will remain after the recession finally recedes and state revenues pick up.
Then what? How should we determine how large a prison population is “right”?
One danger is that we may all get drawn into a debate that is much too narrow. The question of more versus less imprisonment emphasizes the division between those who worry about crime and those who worry about the costs of controlling crime; and it distracts from areas of potential agreement that arise when the focus instead is on the full range of policy choices that affect the crime rate.
If the primary purpose of imprisonment is indeed crime control, then what are the alternatives and what are their social costs? Are there ways to re-allocate our society's resources to reduce the burden on society from both crime and crime control?
With these questions in mind, we have (together with Justin McCrary of Berkeley Law School) organized a new Working Group on the Economics of Crime under the auspices of the National Bureau of Economic Research (NBER). One of our first products is an edited volume from the University of Chicago Press called Controlling Crime: Strategies and Tradeoffs.
Our authors, a group of distinguished economists and other social scientists, review concepts and findings relevant to crime control in a number of domains, including the courts, schools, mental health, income transfer programs, and business improvement districts, with a particular emphasis on paying attention to results from randomized experiments or “natural experiments.”
The overriding conclusion is that there is “money on the table:” crime could be controlled at the current level at lower financial and human cost. Here we sketch our conceptual framework for approaching crime control, and highlight some of the most interesting empirical conclusions.
Conceptualizing Crime and Crime Control
Our vision began with the view that crime is a complex, multifaceted phenomenon for which the most productive portfolio of responses is likely to be quite diverse.
Traditionally, much of the work by psychologists has focused on understanding which individual attributes put some people at elevated risk for criminal offending, while most of the research by economists has focused on the role played by environmental conditions. A more helpful way to imagine the optimal crime-control portfolio is to think of observed crime rates as the outcomes of an interaction between potential criminals (which is to say, most everyone), and the environment of opportunities, licit and illicit, for achieving individual goals.
In this account, the promising crime-reducing interventions include both those that focus on changing the structure of opportunities, and those that invest in individuals to improve their access to licit opportunities while strengthening their resistance to criminal enticements.
Much of the public conversation about crime has focused on just one aspect of this framework, which has also been the focus of much of the psychology literature–the character of the youths.
In the simplistic version, the population consists of good guys and bad guys. The bad guys commit crimes and the good guys do not. The crime rate is proportional to the number of bad guys who are at large. Crime control then is a matter of locking up as many bad guys as possible (or, when the bad guys are immigrants, deporting them).
If offenders are made, not born, the number of “bad guys” can also be reduced through better parenting. In any event, the natural tendency is to seek explanations for crime increases in the character of the youths, and some version of the old refrain of “What's the matter with kids today?”
What is missing from the “good guys, bad guys” account of crime is that crime is a choice, and as such is influenced by incentives as well as character.
Indeed, the theory of crime developed by economists begins where the “character” analysis leaves off, and focuses on how incentives influence crime-related choices for someone of given character. By focusing on crimes as choices made in response to the available options and individual circumstances, the economists' framework provides guidance in understanding trends and patterns, and also in identifying some of the interventions that are likely to be effective in reducing crime. The goal can be loosely expressed as taking the profit out of crime, which invites discussions on such topics as how to incentivize installation and use of better locks and alarms, how to shrink illicit markets for drugs or stolen property, and how best to provide at-risk youths with attractive licit opportunities.
Much of the relevant action is in the private sector.
For violent crime, it is useful to understand that “criminal opportunity” is influenced by social context. Youths typically commit their crimes in groups, and organized criminal gangs are contributors to crime and violence in some cities.
Curtailing criminal opportunity is an important goal for much of the crime-control portfolio of interventions, but not the whole story. The choice to commit crimes also reflects the opportunity cost of crime involvement, which is to say the quality of licit opportunities, as well as personal qualities encapsulated in the term “personality.” Making well-directed investments in child development, and in skill building at all ages, can shape character, and be part of the crime-control portfolio.
There are a variety of interventions that may help direct individuals away from a crime trajectory, starting at birth or even before. A common feature of these efforts is that these programs supplement and support the traditional role of the family in providing for children and instilling human and social capital.
Our findings can be organized into two large and somewhat overlapping bins, “changing the offending environment” and “changing individual propensities towards crime.” Here we recap the list of topics and identify some of the programs that appear particularly promising.
Changing the offending environment
1. Swiftness and certainty, not severity, of punishment: Much of the increase in prison population comes from an increase in average sentence lengths. Yet new data from the Hawaii HOPE experiment found that frequent drug testing followed immediately by a very short jail stay for dirty urine was successful in reducing drug use and criminality among probationers.
Studies of the federal government's COPS program provide further empirical support for the growing suspicion that swiftness and certainty of punishment may actually be most important for controlling crime. The notion that crime is reduced by simply putting more police on the streets without changing what they do, and that deterrence (rather than simply incapacitation) may be an important mechanism behind this result, overturns the conventional wisdom that prevails in many criminology circles.
2. Demand curves for criminogenic goods are negatively sloped: The federal and state excise tax rates on beer and liquor have declined markedly during the last 50 years. The empirical evidence that raising taxes and prices would reduce some types of crime is very strong.
3. Private co-production: Most of the criminology literature around crime control strategies has focused on the role played by government and non-profit interventions. Private citizens and businesses account for a surprisingly large share of all the time and money devoted to preventing crime. State and local governments can help reduce crime by encouraging private action that makes law enforcement more productive.
Two examples for which the benefits exceed costs by an order of magnitude are: creation of the police-tracking infrastructure for LoJack; and creation of the legal framework that facilitates successful Business Improvement Districts (where local businesses are subject to mandatory tax payments that go towards things like hiring extra private security to patrol the area).
Changing individual propensities towards crime
1. The difficulties of changing poverty and adverse mental health: While theory has emphasized the role of economic disadvantage and mental health problems in contributing to criminal behavior, evidence suggests that job training and mental health courts are not the most cost-effective ways to control crime–not because these disadvantages don't matter, but because they are so difficult to modify in practice.
2. Compulsory schooling age: The average high school graduation rate in the America's 50 biggest urban school systems is about 50 percent. One of the few levers available to policymakers to get youth to stay in school is to raise the compulsory schooling age, although it is natural to wonder what good schooling will do for youth who are being forced to go against their will. It is thus striking that we have strong quasi-experimental evidence that cohorts exposed to an increased compulsory schooling age have reduced crime involvement.
3. Social-cognitive skill interventions: Most of the economics-of-crime literature has focused on ways of reducing crime by changing the incentives that confront potential offenders, with very little attention to whether there are ways of helping people respond more productively to the incentives they already face.
There is a growing body of correlational evidence showing that social-cognitive skills that influence how people respond to incentives, such as impulse control, inter-personal skills and future orientation, are predictive of criminal involvement, schooling, and employment.
What the status quo costs us
Our review of the best available social science suggests that the current U.S. approach to crime control is inefficient – more crime control could be accomplished with the same level of resources.
Imagine that we changed sentencing policies and practices in the U.S. so that average prison sentence lengths reverted to 1984 levels. We estimate that this would reduce our prison population by around 400,000 people and reduce total prison spending (currently equal to $70 billion annually) by about $12 billion.
What would we give up by reducing average sentence lengths back to 1984 levels? In terms of crime control, the answer may be: not all that much.
What could we do with this $12 billion in freed-up prison spending? One possibility would be to put more police on the streets. Currently the U.S. spends around $100 billion per year on police protection, so this hypothetical policy switch would increase the police budget by 12 percent and put perhaps as many as 100,000 more police officers on the streets.
This resource reallocation would lead to a decline of hundreds of thousands of violent and property crime victimizations each year. A different way to think about the potential size of the efficiency gain here is to note that the benefit-cost ratio for increased spending on police may be on the order of 4 to 1. If the benefit-cost ratio for marginal spending on long prison sentences is no more than 1 to 1, then reducing average sentence lengths to 1984 levels in order to increase spending on police could generate net benefits to society on the order of $36 billion to $90 billion per year.
Suppose we devoted the resources freed up from a $12 billion cut in prison spending towards Head Start. This 17 percent cut in the prison budget would support a 150 percent increase in the annual Head Start budget (currently around $8 billion per year). Current Head Start funding levels are enough to enroll only around half of poor 3- and 4-year-olds in the U.S.
A 150 percent increase in Head Start's budget could dramatically expand the program on both the extensive and intensive margins. A defensible guess is that re-allocating resources from long prison sentences to early childhood education might generate between $12 billion and $60 billion in net benefits to society.
We might do even better still by focusing on human capital investments in the highest-risk subset of the population–namely, trying to address social-cognitive skill deficits of young people already involved in the criminal justice system.
Research has found that a small fraction of each cohort commits the bulk of all crime. While early childhood interventions have the benefit of targeting people during the time of life in which they may be most developmentally “plastic,” interventions directed at adolescents and young adults enable us to more tightly target those who have emerged as the most likely members of that high-offending subset through their arrest histories.
What sort of social-cognitive skill development could we provide to high-risk young people with $12 billion per year? With around $1 billion annually, we could provide functional family therapy (FFT) to each of the roughly 300,000 youths on juvenile probation each year.
FFT costs approximately $2,500 per youth, with a benefit-cost ratio that may be as high as 25 to 1 from crime reductions alone. With the remaining $11 billion we could provide multi-systemic therapy (MST) to almost every person age 19 and under who is arrested each year.
Diverting $12 billion from long prison sentences to addressing social-cognitive skill deficits among high risk youth could generate net social benefits on the order of $70 billion per year.
Our estimates provide a suggestion of the efficiency gains that could result from reallocating resources from prison to other uses that will, among other outcomes, reduce crime.
A key challenge we face is that our government systems are not well suited to converting the fifth year of a convicted car thief's prison term into an extra year or two of Head Start for a poor child. Government agency heads have strong incentives to maximize the budgets of their agencies, and pour any resources that are freed-up from eliminating ineffective program activities back into their own agencies.
This is the intrinsic difficulty of rationalizing policies across domains, agencies and levels of government. If we could solve this problem–and orient the policy system to up-weight evidence from design-driven research–then, in our quest for effective crime control, it appears possible that we could have more for less.
Philip J. Cook is the ITT/Sanford Professor of Public Policy, Economics, and Sociology at Duke University. Jens Ludwig is the McCormick Foundation Professor of Social Service Administration, Law, and Public Policy at the University of Chicago. A version of this paper appeared in The Criminologist, the newsletter of the American Society of Criminology. They welcome comments from readers.