This article analyzes a general equilibrium model in which agents choose to
specialize in either legitimate or criminal activities. Expenditures on po
lice to apprehend criminals, as well as income redistribution, are determin
ed endogenously through majority voting. We investigate how crime, redistri
bution, and police expenditures depend on characteristics of the underlying
distribution of income-earning abilities and on the apprehension technolog
y. Our model accounts for the positive correlation between inequality and c
rime, the positive correlation between expenditures on police and redistrib
ution, and the lack of correlation between crime and redistribution observe
d in U.S. data.