The effectiveness of criminal sanctions: a natural experiment

Criminology Research Council grant ; (33/96-7)

This study examined data from New South Wales local courts between 1992 and 1997. Using a natural experimental design, it compared cohorts of offenders appearing before magistrates within the 21 courts where random allocation of offenders was used. Variations in sentencing mix between magistrates within each court provided the basis for the analysis. In general, sanctions made very little difference to reoffending rates. However prison had a detectable influence on reoffending for more serious offenders (for example, offenders convicted of burglary or vehicle theft with a prior record), increasing reoffending rates by several percentage points relative to community sanctions.

There was also an apparent "incapacitation effect" resulting from being incarcerated for more than six months; offenders make up for their lost offending opportunities within three years of sentence. For the least serious offenders (such as those convicted on one count of using cannabis), bonds or dismissals reduced reoffending levels compared to fines.

In the middle range of sanctions, the impact of fines and community sanctions were similar. Despite various limitations of the data, the study provided an insight into the small but useful ways sentencing policies can contribute to a reduction of crime in the community.