This paper has attempted to understand how socio-economic factors influence crime in South
Africa at the police precinct level. We find that the relationships between crime and these
variables are influenced by the type of crime (property, robbery, or violent) and the precinctlevel
unit of analysis. We find that precinct-level income and inequality are strong signals of
returns to property crime. Specifically, we find support for our protection hypothesis, which
posits that crime first increases and then decreases with these measures of income and
inequality. We argue that crime increases with income, since this represents returns to crime,and crime increases with inequality, since inequality signals a relative increase in returns to
crime within that precinct. However, at some level of income and inequality individuals become
aware of their vulnerability to acts of crime and take measures to protect themselves, leading
to a downturn in property crime.
Protection aside then, there is evidence of a positive relationship between precinct-level
income and inequality, and property crime. This finding is supported by Demombynes and
Özler (2005), who urge policy-makers to pay attention to precinct-level inequality.
Demombynes and Özler (2005) also explain that such a positive relationship with income does
not imply that police resource allocation should categorically increase with income, since this
would be retrogressive, as it is precisely the wealthy that can afford private protection. Indeed,
reports from the Khayelitsha Commission have already laid bare the destructive effect of
leaving poorer areas under-resourced in favour of wealthier ones
Although robbery crime falls under the bracket of resource acquisition crime, it appears to
have relationships with socio-economic variables that are distinct from those of property crime.
In our preferred regression, no socio-economic variable was found to be statistically
significantly linked to returns to robbery crime.
