Abstract
This study investigates the possible determinants of selected reported crimes in New Zealand from 1962 to 2001. Using annual time series data, six offence groups are individually examined, namely homicide, drugs (cannabis), drugs (not cannabis), sex attacks, robbery and burglary. Previous literature indicates that several factors can influence the levels of crime in a nation. They are referred to generally as deterrence, economic, social and demographic factors. A stepwise backward elimination approach (using Multiple Regression Analysis) and General-to-Specific modelling approach (using PcGets) are applied to the general model, which includes a range of potential determinants. Additionally, models are estimated using two-stage least squares, to control for any simultaneity bias between crime and the explanatory variables. The results suggest that the four groups of factors referred to also provide the main determinants of crime in New Zealand. In particular, the previous year's crime rate, the conviction rate and severity of punishment, the proportion of total births that are ex-nuptial, and the proportion of youth in the population have a significant influence across a majority of the crimes investigated. However, wage, unemployment, divorce and social welfare payments also have a significant effect on some of the offence groups examined in this study.