Steep decline in ex-offenders' probability of re-offending

After about 7.7 years after their last offense, if new Justice Department-funded research is correct, people with robbery convictions not only are much less likely to commit a new offense but thereafter are statistically less likely than the general population to commit new crimes. (Click on the chart to enlarge.) For burglars and those convicted of aggravated assault, the thresholds are even earlier (3.8 and 4.3 years, respectively).

This from a new reentry study focused on post-conviction employment I noticed via Doc Berman and Ted Gest. According to USA Today ("Study could ease concern over hiring ex-offenders," July 1):

A study funded by the Justice Department concludes that over time accused robbers, burglars and batterers pose no greater risk to employers than job candidates in the general population.

In a review of 88,000 arrestees in New York state, Carnegie Mellon University investigators found, for example, that after about 7 1/2 years the "hazard rate" for an 18-year-old first-time arrestee for robbery declined to the same rate as an 18-year-old in the general population. For 18-year-olds arrested for aggravated assault, it took about four years to reduce the risk.

Hazard rates are calculated based on the time the suspect remains free from re-arrest. The calculation also accounts for the fact that risk of arrest generally declines with age.

"We believe that our analysis provides the criminal justice community with the first scientific method for estimating how long is long enough for someone with a prior record" to no longer be considered a special risk, according to the study authored by Carnegie Mellon criminologist Alfred Blumstein. ...

The study focused on three offenses — robbery, burglary and aggravated assault — because they represented some of the largest sample numbers.

According to the National Institute of Justice, which created the graphic at the top of this post with data from Blumstein's study:

The probability of new arrests for first offenders declines with time from first arrest and eventually becomes lower than that of the general population. For those in the general population who were first arrested in 1980, the probability of being re-arrested decreased steadily the longer they stayed clean of further involvement with the criminal justice system. They can be compared to the general population, which mostly includes people never before arrested, as well as to those recently released from prison, who have a high risk of re-arrest. The probability of re-arrest of the 1980 arrestees who stayed clean eventually dropped below that of other people of the same age. For those first arrested for burglary at age 18 years, in 1980, the crossover occurred 3.8 years later at age 21.8. If their first arrest was for aggravated assault, crossover occurred 4.3 years later at age 22.3, and if the arrest was for robbery, it was 7.7 years later at age 25.7. The probability of re-arrest at each of these crossover points was slightly less than 10 percent.