AIM: To see whether diversion of minor drug offenders from criminal prosecution reduces the risk of further offending.

METHODS: We exploit the staged roll-out of a policy implemented in the State of New South Wales (Australia) in 2000, which gave police the discretion to formally caution rather than charge adults detected for use and/or possession of cannabis. We use the exogenous variation in the timing of the policy implementation as an instrument in a 2SLS (linear probability model) analysis to identify the effect of cannabis cautioning on risk of re-offending. A series of checks are carried out to test the robustness of the findings obtained via the linear probability model.

RESULTS: Being given a cannabis caution rather than being prosecuted for cannabis possession reduces recidivism at 5 years by about 16 percentage points, on average. The effects are larger for those under the age of 30, those who have never previously been imprisoned, and those who have already had one caution.

CONCLUSIONS: The adoption of alternatives to prosecution for minor drug offences is more likely to enhance public safety rather than reduce it.