The Bureau of Justice Statistics published a study in 2018 tracking a sample of people released from state prisons in 2005. Within three years, 67.8 percent had been rearrested. Within nine years, that figure climbed to 83 percent. These numbers circulate widely in debates about criminal justice. Depending on who is citing them, they are evidence that rehabilitation does not work, that the people involved are irredeemably dangerous, or that we are doing something systematically wrong.
The third interpretation is correct, but understanding why requires looking carefully at what rearrest actually measures.
Rearrest is not the same as reconviction. The BJS study found that 44 percent of the sample was reconvicted within three years, significantly lower than the rearrest rate. It is not the same as returning to prison. About 37 percent of the sample returned to prison within five years. Some of those returns were for new crimes. Others were for technical violations: failing a drug test, missing a probation meeting, being found outside an approved area. Technical violations now account for roughly a quarter of all prison admissions nationally, up from about 17 percent in 1980.
In other words, a significant portion of the "recidivism" that shows up in these statistics is the criminal justice system recycling people for administrative infractions, not for new crimes that harmed anyone.
The rearrest numbers also cluster heavily in the first year after release. Nearly half of those who were rearrested within nine years had their first rearrest within the first twelve months. This pattern is not a mystery. The period immediately after release is when people face the sharpest combination of destabilization: no housing, no income, no medications, severed family ties, and the concrete barriers of a criminal record that close off employment and housing options. The system releases people into this environment with, in many states, $40 and a bus ticket.
The relevant policy question is not whether recidivism is high. It is why it clusters in the first year. When researchers look at interventions that address the specific risks concentrated in that window, the results are dramatic. Pre-release planning that arranges housing and employment before release, transitional jobs programs that provide immediate income and structure, medication continuity that prevents psychiatric crises, and community mentorship programs that rebuild social networks all reduce the 12-month rearrest rate substantially.
The Honest Opportunity Probation with Enforcement initiative in Hawaii found that consistent, immediate, brief sanctions for probation violations combined with treatment access reduced rearrest rates by 50 percent compared to standard probation. Randomized controlled trials of transitional employment programs have found effects of 30 to 40 percent on one-year recidivism. These are not marginal improvements.
The 67 percent figure is real. It also reflects a system that takes people, strips away their support structures, and releases them with inadequate re-entry resources into communities that have been told they are dangerous. The number is telling us something about what happens when you design a system this way. We could design it differently.