We revisit the association between property crime and income and inequality using data from South Africa in 2011. We investigate the possibility that income and inequality vary non-linearly with property crime, based on a strong theoretical foundation about investment in private security. We use a combination of 2011 South African census data and reported burglary rates from the South African Police Service (SAPS) at the level of 853 police precincts (or aggregations thereof) covering the whole country. To test for non-linearity between our variables of interest, we take advantage
of the flexibility of a Generalised Additive Model (GAM), which has the additional advantage of being able to control for the spatial correlation of crime. We find evidence of an inverted U-shaped relationship between burglary and local area income; which we rationalise as the result of richer people investing in private security. We also find that when income is low, a unit increase in inequality leads to a decrease in crime; but when income is high, a unit increase in inequality leads to an increase in crime. We think this interaction is related to how local inequality mediates the gulf between resources available to local elites to protect themselves, versus local lower-income individuals to commit crime. When areas are, in fact, very poor and very unequal, more inequality might represent a simply insurmountable gap between the ability of elites to invest in sophisticated private security and the ability of lower income criminals to break and enter.