August 2023
First Step Act
An Early Analysis of Recidivism
Analysis by Avinash Bhati, PhD
Passed in 2018, the First Step Act (FSA) was designed to reduce reoffending among people leaving federal prisons. Provisions of the FSA are summarized in a Congressional Research Service brief.
According to data published by the federal Bureau of Prisons (BOP), 29,946 people were released from BOP facilities under the FSA from 2020 to January 2023.1
This analysis estimates recidivism rates among individuals released from BOP prior to the FSA who had similar risk profiles and were tracked for similar periods of time as those released under the FSA. Without access to data on individuals, researchers made several plausible assumptions in order to produce estimates using aggregate data that has been made public by the BOP and Office of the Attorney General (OAG). Those assumptions are explained in a supplemental methodology report. The analysis does not seek to evaluate the FSA or its implementation.
Recidivism Rates
Recidivism rates were 37% lower for people released under the FSA.
The BOP defines recidivism as any rearrest or return to federal prison for a new crime or technical violation of supervision, regardless of the outcome of that arrest (whether a person is charged or convicted). According to BOP's published data, the recidivism rate for all people released under the FSA is 12.4%.2 While that is considerably lower than the 46.2% recidivism rate3 for all people released from BOP facilities in 2018, people released under the FSA differ in important ways from all persons released from federal prisons, including their assessed risk level and the length of time they have been out of prison. To account for these differences, this analysis estimates recidivism among people released prior to the implementation of the FSA who are similar to people released under the FSA in two key ways: (1) they had a similar risk assessment classification according to BOP’s current risk assessment tool, and (2) they had been released to the community for a similar amount of time.
The analysis shows that people released prior to the FSA who were at similar risk of reoffending and had similar amounts of time in the community had an estimated recidivism rate of 19.8%. This means the recidivism rate for people released under the FSA is roughly 37% lower than similarly situated people released prior to the FSA.4 The analysis also found lower rates of recidivism among people released under the FSA within all risk levels as assessed by BOP's risk assessment tool.
Arrests Incurred
People released under the FSA incurred 3,125 fewer arrests.
Drawing on the recidivism analysis above, it is possible to estimate the total number of arrests (for new crime or technical violations) incurred by people released under the FSA and by 29,946 similarly situated people (i.e., risk level and time since release from prison) released from federal prisons prior to the FSA. With a recidivism rate of 12.4%, the 29,946 people released under the FSA over three years could have accounted for between 3,712 and 4,330 arrests in the community over three years. With an estimated recidivism rate of 19.8%, an equal number of similarly situated people released prior to the FSA could have accounted for between 5,918 and 7,455 arrests in the community over three years. Taken together, these estimates suggest that people released under the FSA could have accounted for between 2,207 and 3,125 fewer arrests in the community over three years when compared to similarly situated people who were released from federal prisons prior to the FSA.
National Arrests
People released under the FSA accounted for 0.02% of national arrests.
While the number of arrests in the U. S. varies over time, from 2020 to 2022, there were approximately 20 million arrests made nationally.5 The highest estimated number of arrests incurred by people released under the FSA (4,330) would account for two one-hundredths of one percent (0.021%) of all arrests nationwide during that time period, or 21 out of every 100,000 arrests. The highest estimated number of incurred arrests for similarly situated people (i.e., risk level and time since release from prison) released from federal prisons prior to the FSA would account for 0.037% of all arrests, or 37 out of every 100,000.
0.02%
Conclusions, Limitations, & Next Steps
The analysis suggests that people released under the FSA have lower recidivism rates than rates estimated for similarly situated people (i.e., risk level and time since release from prison) released from federal prisons prior to the Act's implementation. The analysis does not imply that this reduction was caused by the FSA alone. While there is no clear, causal explanation for this reduction, several factors, individually or collectively, may have led to lower recidivism among people released under the FSA:
Roughly 54% of the individuals released under the FSA completed at least one evidence-based recidivism reduction program. It is possible that these programs successfully reduced recidivism among participants. It is also possible that people with the most motivation for post-release success were more likely to complete these programs and therefore may have had lower rates of recidivism regardless of program participation. However, BOP acknowledges that implementation of the programs was slowed due to the COVID-19 pandemic and other factors. Therefore, it is unlikely that the direct causal impact of FSA programming would be this large (i.e., a 37% reduction).
People released under the FSA were released from federal prisons at varying times from 2020 to 2022 and were tracked for recidivism after their release. The individuals released prior to the FSA who were included in this analysis were all released from prison in 2018 and tracked for recidivism from 2019 to 2021. The different coverage periods for members of the two groups may influence the observed and estimated recidivism rates. In particular, disruptions in police, court, and corrections operations caused by the pandemic may explain some of the differences in recidivism rates.
The analysis above used BOP’s risk assessment tool to identify individuals with similar risk profiles who were released prior to the FSA. It is very possible that, even within each risk level, people released under the FSA are different in important ways from those released prior to the FSA. This means the two groups may differ in ways not accounted for by risk level alone, which may explain the difference in recidivism rates.
A more comprehensive comparison between people released under the FSA and those released under other mechanisms is necessary to properly evaluate the impact of the FSA on recidivism and crime. Detailed individual-level data about the people released and under which mechanism they were released, whether they participated in evidence-based recidivism reduction programs or productive activities during custody, the type of recidivism (violent, property, drug, other offense, or technical supervision violation), and how good time and earned time sentence credits impacted the length of their prison term is needed to inform such a comparison.
While this analysis attempts to construct a more valid comparison of people released under the FSA and people released from BOP prior to the FSA, its findings should not be interpreted as an impact assessment of the FSA. Undertaking a well-designed, individual-level study is an important next step in evaluating the impact of the legislation.
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.