This study assesses the extent to which the probability of arrest, the probability of imprisonment and imprisonment duration impact on property and violent crime rates in New South Wales.
Method: A dynamic panel data model with fixed Local Government Area and time effects was adopted to explore this, while adjusting for potential confounders of the relationship between arrest, imprisonment and crime. The first-differenced generalised method of moments was used to estimate the model parameters.
Results: One per cent increases in arrest rates for property and violent crime are estimated to produce 0.10 per cent and 0.19 per cent decreases in property and violent crime, respectively. If the one per cent increase in arrest rates is sustained, the long-run effect is estimated to be 0.14 and 0.30 per cent decreases for property and violent crime, respectively. The short-run elasticities for imprisonment probabilities were smaller, (-0.09 and -0.11), as were the long-run elasticities (-0.12 and -0.17), for property and violent crime, respectively. There was no evidence that increases in the length of imprisonment has any short or long-run impact on crime rates.
Conclusion: The criminal justice system plays a significant role in preventing crime. Some criminal justice variables, however, exert much stronger effects than others. Increasing arrest rates is likely to have the largest impact, followed by increasing the likelihood of receiving a prison sentence. Increasing the length of stay in prison beyond current levels does not appear to impact on the crime rate after accounting for increases in arrest and imprisonment likelihood. Policy makers should focus more attention on strategies that increase the risk of arrest and less on strategies that increase the severity of punishment.
Authored by Wai-Yin Wan, Steve Moffatt, Craig Jones and Don Weatherburn.
Image: 'Chock-a-block: state's jails bursting at seams', publik16 / flickr