Income Polarization and Crime: A Generalized Index and Evidence from Panel Data

Abstract

The empirical evidence for a positive relationship between income inequality and crime is mixed at best. This paper questions the use of standard income inequality measures (e.g., Gini coefficient) in this literature and makes the case that polarization of the income distribution is a crucial aspect in understanding the incentives for criminal activities. Two hypotheses are presented: (i) for a given level of inequality, polarization reduces between-group income mobility and thus raises crime incentives; (ii) for a given level of between-group income mobility, the poor feel more alienated against the rich than the latter against the former. We test these hypotheses using a new measure of social antagonism based on income bipolarization that allows for asymmetric feelings of alienation between low- and high-income groups. The asymptotic distribution of the proposed bipolarization index is derived and an easy-to-implement jackknife variance estimation algorithm is developed. Fixed effects estimates suggest that when both the Gini and the bipolarization indices are included in the same regression model, bipolarization significantly raises the crime rate, whereas the Gini index does not. It is also found that the proposed generalized bipolarization index has more explanatory power when more weight is placed on the feeling of alienation of the lower income class.