Abstract
Hale and Sabbagh (1991) failed to find cointegration (i.e. a long-term equilibrium relationship) between crime and unemployment in the case of England and Wales. The present study (basedon multiple cointegrationtests preceded by various unit root or non-stationarity tests) is the first attempt at putting the analysis of crime in a temporal Granger-causal framework in the Australian case (1963 - 90) by binding the relationship between different categories of crime and their socioeconomic determinants within a multivariate cointegrated system. The results, based on the most recent methodology, broadly indicate that, although the relative importance of the determinants of crime varied by type of crime, of all the determinants it is 'dwelling commencements' (a proxy for 'wealth') that appears to impact more or less on all the categories of crime significantly, followed by 'urbanization'.