Policy Paper (Interest) 12

The paper examines the association between various crime series for South Africa and their economic determinants between 1960 and 1993. The study uses Johansen cointegration techniques to establish the importance of income per capita (economic opportunities), police officers per capita, conviction rates (both of which proxy for the likelihood of criminal success whilst the latter also captures an institutional efficiency aspect), and political instability (necessary to reflect the unique South African environment during this time) on per capita crime levels. The results caution previous studies that find a close relationship between the level of economic activity and recorded crime. Previous work has relied on the Engle-Granger technique assuming a unique cointegrating vector. Our results for South Africa indicate that whilst total offences are negatively associated with income per capita, disaggregated crime series do not always yield cointegrating vectors. Some conclusions on the policy implications also follow.