The dependent variable in the analysis was the annual rate of robbery (per 100,000 population) recorded by New South Wales police between 1966 and 2000. Since there was no reliable measure of the actual number of dependent heroin users in New South Wales over the study period, the analysis used a proxy variable in the model. The annual rate of deaths from heroin overdose (per million males aged 15-44) was calculated against the robbery rate (per 100,000 population) for the years 1966 to 2000. Both the robbery series and the heroin overdose deaths had an upward trend over the study period. Since the late 1980's, the relative magnitude of the two series has been fairly stable, so a multivariate linear relationship between the series was modeled, incorporating selected control variables, namely, increasing unemployment rates for males, a decreased likelihood of apprehension by police for robbery crimes, and a reduction in the likelihood of imprisonment for robbery. Assuming that there is a linear relationship between the number of dependent heroin users and the rate of heroin overdose, it is possible to calculate the elasticity between robbery and dependent heroin use. On this basis, the study estimated that since 1966 each 10-percent increase in the annual number of dependent heroin users has led to a 6-percent increase in the New South Wales robbery rate. By comparison, each 10-percent decrease in the robbery clearance or imprisonment rate has led to a 3-4-percent increase in robbery. Brief comments are offered on the implications of these findings for policymaking in countering both heroin use and robbery. 4 figures, 13 tables, 4 notes, and 32 references