Responsible researcher: Eduarda Miller de Figueiredo
Author: Evan K. Rose
Intervention Location: North Carolina
Sample Size: 1,603,713 individuals
Sector: Public Security
Variable of Main Interest: Violation of Probation
Type of Intervention: Probation
Methodology: 2SLS, Differences in Differences
Summary
This article studies the primary way the U.S. Criminal Justice System gives convicted offenders a second chance to avoid prison and return to work: parole. Using a 2011 reform in North Carolina, it was studied whether probation violations carry severe punishments to deter recidivism. Through estimations using robust methods, it was noticed that the reform reduced racial differences in the revocation of paroles due to rule violations, however, there was no reduction in racial disparities in prisons.
Recent research has studied racial disparities in decisions made by police, judges, prosecutors, and juries and how arrest, conviction, and incarceration affect economic outcomes (Chetty et al., 2020). However, less attention has been paid to the impact of supervision, which is the most common punishment in the United States.
Every year, more than 4.4 million convicted offenders are sent home under supervision on the condition that they obey strict rules such as no alcohol and drugs, frequent meetings with a social worker, and payment of fines and fees. Failure to comply with these rules may result in incarceration. According to the Justice Center of the Council of State Governments (2019) [1] , supervised offenders are just as likely to be incarcerated for such “technical violations” as for new national criminal offenses, which are most concentrated among black men. .
These technical rules, however, are the main tools that the system uses to monitor supervised offenders and support their reintegration (Piehl and LoBuglio, 2005). Despite the costs, punishing technical rule violations with incarceration – or revoking supervision – can be effective if the violations are strong indicators of future criminal behavior, making them good indicators of the risk of recidivism.
Therefore, this article examines the effectiveness and equity of probation revocation, which accounts for 80% of the supervised population.
In recent decades, the supervised population has grown in parallel with incarceration rates. Compared to 1980 levels, there was an increase of more than 300% and more than double the number of people currently incarcerated. For much of the past 25 years, the state of North Carolina has operated a very small parole system, choosing to release most incarcerated individuals without supervision. However, the analysis in this study focuses only on the supervised probation system.
Probation in the United States is most commonly used for juveniles and first-time offenders, who are facing their first criminal case. In which, in North Carolina, 78% of first-time offenders are placed on probation, along with 70% of offenders ages 16 to 25. Probation periods generally last between 1 and 3 years, where failure to comply leads to the risk of incarceration. North Carolina includes a set of standard rules in probation conditions:
Furthermore, racial disparities are a pervasive feature of the U.S. Criminal Justice System. Where black men who have not completed high school are almost as likely to be incarcerated as they are at work and are employed at half the rate of white men with similar education.
In 2011, North Carolina made changes to the State Criminal Justice System by passing the Justice Reinvestment Act (JRA). It introduced strong limits on the courts' authority to revoke probation, so after 2011 supervision could only be revoked because of the occurrence of new crimes or escape from supervision. Previously, judges could revoke for any technical violation.
Administrative datasets provided by the North Carolina Department of Public Safety (DPS) [2] and all criminal courts of cases resolved from 2006 to present were analyzed. The core data consists of records of individuals serving supervised probation between 2006 and 2018.
The control and treatment groups are young, with 50% of the sample aged 30 or younger at the beginning of the period, the majority male, and minorities overrepresented in relation to the state population. On average, supervised probation periods last about 20 months and are the result of a relatively balanced mix of felony, misdemeanor, and driving (drunk or revoked license).
The treated sample has very limited criminal histories, with the median defendant having only one prior misdemeanor conviction and no prior sentences for supervised probation or incarceration. The majority of probation violations committed were for failure to pay fees and fines. The other most common violations are: failure to report to a probation officer, drug violations, and drug treatment program failures.
In the analysis, a binary dependent variable was used, which will take the value 1 if an offender is arrested for a new crime. To investigate racial differences, three measures of repeal effectiveness were used. The first is predictive accuracy, where when it is high, a large fraction of those revoked will reoffend, whereas when it is close to the population average, revocation would not have a signal value for recidivism. The second and third concepts provide alternative measures of effectiveness by reversing this conditional probability to examine Type I and Type II error rates.
Precision and error rates were estimated using instrumental variables, which allowed the use of an estimation using the 2SLS model. Furthermore, a difference-in-differences approach was also used to compare outcomes for supervised and unsupervised offenders.
The main results show that the number of offenders on supervised and unsupervised probation did not change discretely around the reform, indicating that the sentencing behavior of judges was not affected. Thus, although probation in general became more lenient after the reform, there is no evidence that judges changed their sentencing behavior or potential offenders changed their crime choices in response.
Results from the difference-in-differences analysis show that the 2011 JRA reform had virtually no effect on the unsupervised offender group. In which there was a decrease in revocations by approximately 6p.p.. As offenders did not see a decline in the revocation of parole supervision, their prison rates evolved smoothly throughout the reform.
Regarding racial disparities, the results demonstrate that black offenders continue to see substantially greater reductions in incarceration, but without different changes in recidivism. The evidence suggests that racial disparities in this scenario do not arise due to racial bias on the part of police, judges, or probation officers, but rather reflect differences in behavior between black and non-black offenders. However, such differences are not easily explained with observable characteristics, suggesting that the behavioral differences between these two groups that generate the differentiated impact of revocations, may be a reflection of other more subtle and contextual factors, such as access to informal credit that could be used to pay fees and fines.
The results show that ostensibly race-neutral policies (imposing common-sense rules) can generate large racial disparities that are not justified by the ultimate goals of the policies. In which correcting disparities due to disparate impacts may be easier than changing the biased behavior of decision makers, because that is a matter of simply changing the rules themselves. Thus, the findings presented by the study provide evidence that such policies are viable and can have significant impacts on racial disparities.
References
Chetty, Raj, Nathaniel Hendren, Maggie R. Jones, and Sonya Porter, “Race and Economic Opportunity in the United States: An Intergenerational Perspective,” Quarterly Journal of Economics , 135 (2020), 711–783.
Council of State Governments Justice Center, “Confined and Costly: Hou Supervision Violations Are Filling Prisons and Burdening Budgets,” CSG Technical Report, 2019.
Piehl, Anne Morrison, and Stefan F. LoBuglio, “Does Supervision Matter,” Ch. 5 105-138, in Prisoner Reentry and Crime in America , Jeremy Travis and Christy Visher, eds. (New York: Cambridge University Press, 2005).
[1] Council of State Governments Justice Center.
[2] North Carolina Department of Public Safety (DPS).
Check the Institution's registration in the e-MEC System here
© 2025 All Rights Reserved
