Measuring unfair (in)equality

This paper shows one way of generalizing the standard framework of inequality measurement to allow for a distinction between fair and unfair inequalities. We introduce the unfairness Lorenz curve and the unfairness Gini, which are generalizations of the standard versions of the Lorenz curve and the Gini. With this more general framework in place, we study the implications of responsibility-sensitive theories of justice for the evaluation of the income distribution in Norway from 1986 to 2005. We find that both the pre-tax and the post-tax income distributions have become less fair in Norway, even though the standard Gini for the pre-tax income distribution has decreased in the same period. Two trends explain this development: the increase in income share of the top percentile and the change in the situation of females in the labor market. The concentration of income at the top of the distribution contributes both to increased unfairness and increased inequality, whereas the increase in females' working hours and level of education primarily contributes to a reduction in inequality. Thus, the latter effect dominates for the standard Gini and the former effect for the unfairness Gini. Furthermore, we find that the increase in post-tax unfairness is even larger than the increase in pre-tax unfairness, which shows that the tax system in Norway contributes less to eliminating unfairness in 2005 than in 1986.