The surprisingly weak link between incarceration and crime

10-29-14sfpThe logic of the “tough on crime” movement holds that punishing people harshly for their offenses — whether violent or nonviolent — is a critical tool to prevent crime. That attitude was the driving force behind criminal justice policy in Oklahoma and across the country for years, and states sent more and more people to prison each year as a result. Although crime has been decreasing steadily since its peak in the early 1990s, the incarceration rate only began dropping slowly in the last 8 years or so.

These national trends, however, mask wide variation in trends among the states. Between 2006 and 2014, California’s imprisonment rate decreased by 27 percent while Arkansas’s increased by 23 percent. Over the same period, the violent crime rate dropped 35 percent  in Virginia but increased 32 percent in South Dakota. In Oklahoma, the imprisonment rate rose by 5 percent while the violent crime rate dropped by 21 percent.

As the chart below shows, most states’ violent crime rates fell significantly between 2006 and 2014, while imprisonment rate changes varied widely. The larger circles represent the states with higher imprisonment rates in 2006. In all, 44 states experienced the nationwide trend toward lower violent crime rates, while only 6 states saw increases in violent crime. In spite of the declining violent crime rate in most states, only 26 states saw their imprisonment rates decline.

Source: Brennan Center analysis of Bureau of Justice Statistics and FBI data

In this chart Oklahoma stands out as the state with the highest initial incarceration rate in 2006 where incarceration is still growing. Other high incarceration states like Texas, Louisiana, and Mississippi have begun to move in the opposite direction; Oklahoma hasn’t. 

The wide variation among the states also suggests that incarceration and crime rates simply aren’t very closely related. In truth, the causes of movements in the crime rate are very poorly understood. A comprehensive evaluation of various popular hypotheses estimated that between 1990 and 1999, zero to 7 percent of the crime decline could be attributed to increased incarceration, and no single factor could account for more than 5 percent of the decline with reasonable certainty.

But although the causes of the crime decline are mysterious, there are clear reasons why the imprisonment rate went down in some states and not others. Recognizing that mass incarceration was creating budgetary problems without clear public safety benefits, many states began reforming their criminal justice systems and saw decreases in their prison populations. Other states, including Oklahoma, responded more tepidly. Although Oklahoma’s Justice Reinvestment Act, passed in 2012, contained many of the same reforms that saw success in other states, the poor implementation of the law weakened its effects, and the prison population continued to grow.

The Oklahoma Legislature restarted reform efforts last year with a task force that produced four successful bills to reduce sentences for minor drug and property crimes, give prosecutors more discretion to charge nonviolent crimes as misdemeanors, and expand alternatives to incarceration. Two ballot questions promoted by the Oklahomans for Criminal Justice Reform coalition would push these reforms further. If passed by voters in November, State Question 780 would remove the possibility of a prison sentence for simple drug possession and State Question 781 would direct the savings from diverting people away from prisons into drug and mental health treatment. Initial polling suggests that the ballot questions have strong support.

The experience in other states tells us that we can stop sending so many people to prison without seeing more crime as a result. This year’s reforms are a good start and should have noticeable effects, especially in the short term. But there is still much work to be done; dealing with long-term prison population growth will require confronting overly harsh punishments for more serious offenders, too. As Governor Fallin’s Justice Reform Task Force meets over the coming months, they should focus on ways to make lasting changes to reduce the prison population. Many other states have already shown that simultaneously reducing crime and imprisonment is possible.

Learn More // Do More


Ryan Gentzler worked at OK Policy from January 2016 until November 2022. He last served as the organization's Reserach Director and oversaw Open Justice Oklahoma. He began at OK Policy as an analyst focusing on criminal justice issues, including sentencing, incarceration, court fines and fees, and pretrial detention. Open Justice Oklahoma grew out of Ryan’s groundbreaking analysis of court records, which was used to inform critical policy debates. A native Nebraskan, he holds a Master of Public Administration degree from the University of Oklahoma and a BA in Institutions and Policy from William Jewell College. He served as an OK Policy Research Fellow in 2014-2015.

5 thoughts on “The surprisingly weak link between incarceration and crime

  1. I say crime rates are dropping everywhere because people are increasingly doing their part to make a more peaceful and a more just world. Parents and schools are doing their part. I also say it’s about time law enforcement professionals also get serious about their role in doing exactly the same thing and take jobs with the Police Departments ONLY if they can prove they are more interested in law adherence in general than rooting out or dreaming up “crimes” around every corner. If these people are angry for reasons independent than the tasks at hand, they should not be hired. The pre-recruitment psychological tests need to better identify people with anger issues, and sad to say they are not working now. These jobs must be reserved ONLY for people who can demonstrate in an evidence based way, their penchant for fairness and accurate assessments of situations. And nothing will ever change the growing rates of false arrests, false imprisonments and blatant police killings… unless police officers care as much about peace and nonviolence in their own lives/world as they say they care about supposedly stopping crime. This is pretty much uncharted territory and there’s no excuses why it should continue to be so when the general population have been so long doing their part.

  2. Ryan, your article contains outright falsehoods. Since CA passed prop47 (identical to SQ780), Property AND violent crime have skyrocketed.
    Their DOC headcount only declined by 5% – and only because prop47 was retroactive (780 isnt). Yet they havent seen one red cent in DOC spending.
    How do you think you’ll see any different result with the identical law here?

  3. California crime rates did not “skyrocket”. They saw a slight uptick in 2015 (Prop47 passed in 2014) but they are still at historic lows:

    This article is accurate in all of the statistics that it discusses and in the larger point that highly punitive incarceration policies are not correlated with reduced crime.

  4. Per the book Freakonomics, abortions in 1970s led to crime decline in 1990s

    “First, let’s start by reviewing the basic facts that support the Donohue-Levitt hypothesis that legalized abortion in the 1970s explains a substantial part of the crime decline in the 1990s:”
    “1) Five states legalized abortion three years before Roe v. Wade. Crime started falling three years earlier in these states, with property crime (done by younger people) falling before violent crime.”

  5. I am responding mostly to the Brennan Center for Justice 142 page report. It has an odd pattern of having appendixes that are more insightful than the text. My best guess is that Oliver Roeder is exceptionally math minded. It seems he cares very much about the charting of the data he gets, but is not doing the hard work of examining the integrity of his data. He knows how to do regression analysis and sensitivity calculations and seems in love with CompStat and the numbers that the software provides him.
    The appendix corrects him and points out that CompStat changes the integrity of a community’s crime statistics. While the change is mostly an improvement, it also creates its own pattern of false data.
    The reports author’s seem not to have heard of the Butterfield effect and reproduce it perfectly. To over simply, when looking at the homicide statistics, a high rate of incarceration imprisons future victims (sic) of crime, and because of that phenomena there is a small reduction in the homicide rate. To say this another way, a lot of homicide victims are criminals being killed by other criminals. It happens somewhat that criminals cheat and anger other criminals and get killed.
    The author of the appendix signals the reader that she is aware of the criminogenic effect of imprisonment even if the principal author is not.
    The author of the appendix is aware that an increase in police officers creates an increase in crime statistics by transforming unreported crime into reported crime. The author of the report ignores that very real effect and does not try to measure it.
    I could not find anywhere in the peer review ring someone who is interviewing criminals and seeking their opinions.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.