Introduction
In Spring 2022, the Council on Criminal Justice (CCJ) launched the Task Force on Long Sentences. The group of 16 experts represents a broad range of experience and perspectives, including crime victims and survivors, formerly incarcerated people, prosecutors, defense attorneys, law enforcement, courts, and corrections. Its mission is to examine how long prison sentences—defined as 10 years or more—affect public safety, crime victims and survivors, incarcerated individuals and their families, communities, and correctional staff, and to develop recommendations to strengthen public safety and advance justice.
This paper was produced with support from Arnold Ventures, the Ford Foundation, Southern Company Foundation, and Stand Together Trust, as well as #StartSmall, the John D. and Catherine T. MacArthur Foundation, and other CCJ general operating contributors.
The analysis presented here was commissioned by the Task Force to examine the relationship between long prison sentences and public safety in one state, Illinois. The project was made possible by a partnership between the Illinois Sentencing Policy Advisory Council, CCJ, and Maxarth LLC, a data analytics firm.
Long sentences are often imposed with an expectation that they will prevent some crime by incapacitating individuals and deterring them from engaging in future crime after release. While there are several other purposes of sentencing, including punishing offenders, restoring victims and survivors, and deterring others from committing crimes, the research question explored in this brief is whether people serving long sentences may be confined beyond their likelihood of engaging in criminal behavior. Incarcerating individuals beyond this point would increase the size of the total prison population—and associated costs—without producing additional benefits for public safety.
To examine this question, Maxarth LLC analyzed detailed arrest history data for people who were released from Illinois prisons between June 2016 and June 2019. For the 1,127 people in this release cohort who had served 10 years of more prior to release, microsimulations were created to estimate the number of arrests averted due to the individuals’ long prison terms.1 The number of arrests averted includes: (a) estimated arrests that did not happen because the person was incarcerated (incapacitation effect), (b) arrests that did not happen during a 30-month post-release tracking period because the person reduced their criminal activity (specific deterrence or rehabilitation effects), and (c) arrests that happened post-release because the person increased their criminal behavior following their incarceration (criminogenic effect). The incapacitation effect was calculated using the length of each person’s prison stay (time served); the specific deterrence and criminogenic effects were calculated in the 30 months following each person’s release date. The analysis accounted for the start of the COVID-19 pandemic in March 2020, which overlapped with the 30-month post-release tracking period for some members of the sample. See the supplemental methodology report for detailed procedures.
To estimate the impact of shortening lengthy prison terms, these microsimulations were then recalculated to estimate the number of arrests averted when each person’s time served was reduced by one, two, and three years and by 10%, 20%, and 30%.2 Any decline in the number of arrests averted in these simulations can be understood as the number of additional arrests resulting from the prison term reduction. For example, if an individual’s 12-year prison sentence was estimated to avert three arrests, and the microsimulation indicates that only two of these three arrests would be averted if they were released a year earlier, then we can say that a one-year reduction in their prison term would create one additional arrest. To estimate how many additional arrests would take place as a result of releasing people serving long sentences from incarceration sooner every year, all results reflect estimates for the 376 people predicted to be released sooner every year (1,127 members of the sample divided by three years equals 376). See the supplemental methodology report for details about how the microsimulations were conducted and how results were annualized.
Key Takeaways
- Modest reductions in the length of long prison stays would likely result in relatively few additional arrests. Estimates show that reducing time served by one, two, or three years for the 376 individuals released annually in Illinois would result in 11 to 37 additional arrests. Reducing time served by 10%, 20%, or 30% would result in an estimated 17 to 63 additional arrests in Illinois.
- The additional arrests likely to result from reductions in time served would constitute a virtually undetectable increase in the annual volume of arrests in the state. The 11 additional arrests from a one-year reduction in time served comprise one one-hundredth of one percent (0.01%) of all 89,173 index crime and drug arrests in Illinois in 2020.3 The estimated 63 arrests expected to occur should time served be reduced by 30% constitute less than one-tenth of one percent (0.07%) of all index crime and drug arrests in Illinois in 2020.
- The additional arrests that may occur when prison stays are shortened are unlikely to be for violent crimes or weapons charges. Estimates show that a one-year reduction in time served would result in two additional arrests for violent crimes, two arrests for property crimes, two arrests for drug crimes, and five arrests for “other” crimes. The analysis estimates that no additional arrests for weapons charges would occur. A 30% reduction in time served is estimated to result in 10 additional arrests for violent crimes, 9 arrests for property crimes, 10 arrests for drug crimes, 3 arrests for weapons charges, and 32 for “other” crimes. The increase in arrests with a 30% reduction in time served would comprise less than one-twentieth of one percent of property and drug crimes (0.02% and 0.03%, respectively) and less than one-tenth of one percent (0.08%) of all arrests for violent crimes.
- No individual is estimated to have more than one additional arrest. Reducing time served by one, two, or three years and by 10%, 20%, or 30% did not result in any of the 1,127 members of the analysis sample being estimated to have more than one additional arrest each. A 30% reduction in time served resulted in no individual predicted to have more than 0.68 additional arrests. Over 90% of the sample would likely have fewer than 0.3 additional arrests, while less than 2% would have more than 0.4 additional arrests.
- Reducing long prison stays can significantly affect the size of the total prison population. Reducing lengthy time served by one year could reduce the average daily prison population in Illinois by 1.3%; reducing lengthy prison stays by 30% would decrease the prison census by 7.2%.
- The relatively small number of additional predicted arrests indicates that the vast majority of the people serving long prison terms have “aged out” of criminal behavior near the end of their stays, and that reducing long prison terms for them has little to no public safety impact. Because these same individuals are serving the longest prison terms, they make up a disproportionate share of the incarcerated population. As such, shortening their prison terms can reduce the size of the prison population without significantly compromising public safety.
Glossary of Terms
- Age-crime curve: The fact that as most people age, their likelihood of engaging in criminal offending first increases and then decreases (referred to as “aging out” of crime)
- Criminogenic effect: Producing or tending to produce crime
- General deterrence: The idea that punishing an individual offender will dissuade others from committing crimes
- Incapacitation: The prevention of crimes in the community that might otherwise be committed if an individual was not incarcerated
- Long sentences: The Task Force on Long Sentences defines long sentences as prison sentences of 10 years or more. Numerous state and federal statutes use 10 years as either the maximum or minimum allowable term of imprisonment
- Rehabilitation: The act of helping people desist from criminal behavior through therapeutic programs, activities, incentives, substance use treatment, and other services designed to change behavior
- Specific deterrence: The idea that punishing an individual offender will dissuade that individual from committing future crimes
- Time served: The actual amount of time a person spends in prison as distinct from the judicial sentence
Finding 1: A modest reduction in the length of long prison stays would likely result in relatively few additional arrests
Figure 1 shows how many additional arrests are predicted to occur based on six different simulations for reducing time served (reducing time served by one, two, or three years and by 10%, 20%, or 30%). Note that these additional arrests are compared against the estimated arrests averted under the actual length of stay for each individual, not relative to a baseline of zero arrests.
Figure 1: Estimated Additional Arrests from Reducing Lengthy Prison Terms
Calculations for this analysis show that the number of additional arrests expected to result from shortening lengthy prison terms is relatively small. For example, a policy that reduces lengthy prison terms by one year for the 376 persons released annually after serving 10 or more years would result in an estimated 11 additional arrests in Illinois. Reducing time served by two or three years results in an estimated 23 or 37 additional arrests, respectively. Reducing overall time served by 10%, 20%, or 30% also implies small increases in estimated arrests. A 10% reduction in time served would result in 17 additional arrests, a 20% reduction would lead to 37 additional arrests, and a 30% reduction would generate 63 additional arrests in Illinois.
Finding 2: The additional arrests likely to result from reductions in time served would constitute a virtually undetectable increase in the overall volume of annual arrests
Any additional arrests are cause for concern, but the number of additional arrests predicted by this analysis is small relative to the volume of overall arrests in Illinois. Figure 2 shows the estimated additional arrests as a percentage of the total arrests in Illinois in 2020. The 11 additional arrests from a one-year reduction in time served comprise one one-hundredth of one percent (0.01%) of the 89,173 index crime and drug arrests made in Illinois in 2020. The estimated 63 arrests likely to occur should time served be reduced by 30% constitute less than one-tenth of one percent (0.07%) of this total arrest number.
Figure 2: Estimated Additional Arrests as a Percentage of All Index Crime and Drug Arrests in Illinois in 2020
Finding 3: Additional arrests that may occur when long prison stays are shortened are unlikely to be for weapons charges or violent crimes
The analysis also produced information on the types of crimes for which people would likely be arrested based on their patterns of pre- and post-incarceration offending. As shown in Figure 3, few of the additional arrests would be for violent crimes (e.g., assault, homicide) or weapons charges.
Figure 3: Estimated Additional Arrests by Crime Category
Notes: Violent crimes include battery, assault, homicide, and other violent crimes. Property crimes include burglary, trespassing, theft, and other crimes. The “other” category captures arrests for disorderly conduct, motor vehicle offenses, and other crimes. See Table 2 in the supplemental methodology report for a detailed breakdown of all charges included in each arrest category.
Calculations show that reducing time served by one year for the 376 people released annually likely would result in two additional arrests for violent crimes, two arrests for property crimes, two arrests for drug crimes, and five arrests for “other” crimes. The analysis estimates that no additional weapons arrests would occur. If time served was reduced by 30%, an estimated 10 additional arrests for violent crimes would occur, along with 9 arrests for property crimes, 10 arrests for drug crimes, 3 arrests for weapons charges, and 32 for “other” crimes.
As depicted in Figure 4, in the case of a one-year reduction in time served, additional arrests in each category comprise less than one one-hundredth of one percent of total arrests in that category. For a 30% reduction in time served, additional arrests for violent crimes make up less than one-tenth of one percent (0.08%) of all arrests in that category, while arrests for property and drug crimes each comprise less than one-twentieth of one percent of the total (0.02% and 0.03%, respectively).
Figure 4: Estimated Additional Arrests by Crime Category as a Percentage of Arrests in that Category in Illinois in 2020
Note: The “weapons” and “other” charge categories are not included here due to data limitations.
Finding 4: No individual is estimated to have more than one additional arrest
The analysis also generated estimates of how the additional arrests might be distributed across the sample and potentially concentrate among specific individuals. Because the microsimulation analysis produces an estimated arrest value individualized to each person, it is not possible to display results in whole numbers or estimate who would experience exactly zero additional arrests. Instead, values closer to zero indicate that an individual has an extremely low risk for any additional arrest; values around one indicate that an individual would likely have one additional arrest; values around two indicate that an individual would likely have two additional arrests, etc.
Figure 5 displays the number of additional arrests estimated to occur for each individual in the sample when prison stays are reduced by 10%, 20%, and 30%. The amount of time over which additional arrests might occur in the 10%, 20%, and 30% simulations varies for each member of the sample as a function of their overall time served and the amount by which time served is reduced.4 Results for the one-, two-, and three-year sentence reduction scenarios are not visualized, as estimates of additional arrests were smaller and followed a similar pattern. Moving from left to right across the figure indicates the number of additional arrests to be expected from increasingly higher contributors to the overall impact of the simulated policy. For example, in the 20% reduction scenario, the plot indicates that 80% of individuals in the sample are estimated to contribute 0.1 additional arrests or fewer each. Across all six simulated models, no individual estimates exceeded 1.0, indicating that none of the 1,127 members of the sample were estimated to increase arrests by more than one should time served be reduced, even by up to 30%. In fact, a 30% reduction in time served resulted in no individual predicted to have more than 0.68 additional arrests. Over 90% of the sample is estimated to have fewer than 0.3 additional arrests each; less than 2% of the sample is estimated to have more than 0.4 additional arrests each.
Figure 5: Estimated Additional Arrests per Person
Finding 5: Reducing long prison stays can significantly affect the size of the total prison population
Reducing time served for people serving long sentences will influence the total size of the prison population. Releasing some individuals sooner naturally means that prison populations will fall. Figures 6 and 7 show how much the Average Daily Population (ADP) in Illinois prisons would decrease if time served was reduced by one, two, or three years, and by 10%, 20%, or 30%. Figure 6 shows this change in raw numbers; Figure 7 presents this value as a percentage of the ADP.
Figure 6: Estimated Decrease in Average Daily Population
These results show that a one-year reduction in lengthy prison terms could reduce the ADP in Illinois by 376 individuals, constituting a 1.3% reduction in the total prison population.5 Similarly, shortening time served by two or three years would yield ADP reductions of 751 (2.6%) and 1,127 (3.9%), respectively. At the highest end of the estimates performed, a 30% reduction in time served would decrease the daily prison census in the state by 2,110 individuals, or 7.2%.
Figure 7: Estimated Decrease as a Percentage of Average Daily Population
Reducing time served—and, by extension, the total prison population—would have financial implications for Illinois. The Illinois Sentencing Policy Advisory Council conducted an analysis using dynamic marginal costs to estimate the fiscal implications of reducing the number of individuals held in state prisons.6 This analysis shows that a 1,000-person reduction from its daily state prison population represents $12.5 million of the funds that the Illinois General Assembly annually appropriates to IDOC, and a 3,000-person reduction represents $147.9 million. As a point of reference, simulations of reducing time served by three years and 20% resulted in decreases of more than 1,000 individuals, while a 30% reduction resulted in a decrease of more than 2,000 individuals. While these reductions in and of themselves would not automatically produce cost savings for Illinois, they could lead legislators to make different choices about how to fund IDOC’s general operations, including allocating the funds currently dedicated to supporting people serving long sentences into both prison- and community-based programs proven to reduce crime and victimization.
Conclusion
Evaluating the utility and effectiveness of long sentences raises important questions, including how reducing such sentences might affect public safety. Findings from microsimulation analyses indicate a tradeoff: All reductions in time served are estimated to result in a small number of additional arrests, but they would likely be a small fraction of all annual arrests in Illinois, and would likely not be for violent crimes or weapons charges. These results suggest that most people serving long sentences have “aged out” of criminal activity—especially violent crime—near the end of their prison stay. Because any additional arrests would likely be for relatively minor crimes, reducing time served would reduce the size—and expenditures associated with—the prison population without jeopardizing public safety. Although results from similar analyses in other states may differ, the findings presented in this brief indicate that shortening lengthy prison terms can be done without significantly compromising public safety.
About the Author
Dr. Avinash Bhati is the founder and CEO of Maxarth LLC, a data science company providing creative solutions to real-world problems and opportunities. Bhati earned a Ph.D. in econometrics from American University in 2001. He is interested in all aspects of data science, including predictive modeling, quasi-experimental evaluations, micro-simulation models, and synthetic data tools. Bhati has developed and validated numerous pretrial and post-adjudication risk assessment instruments.