Poverty, Inequality, and Redistribution Under Lexicographic Social Welfare

Unlike aversion to inequality, aversion to poverty resists formalization in welfare economics. One way to assign normative significance to the poverty line is to allow the welfare measure to exhibit a discrete loss from poverty (DLP) at z. However, the resulting redistribution scheme prioritizes headcount-reducing transfers to the borderline poor over transfers to the very poorest, rendering the DLP measure unattractive. The paper remedies this by transcending the conventional real valued welfare measure. It proposes a lexicographic L*-ordering, where the first rank criterion corresponds to an inequality-based evaluation function, while the second rank criterion corresponds to an evaluation function that exhibits DLP. The redistribution scheme entails transfers to the poorest until the first rank criterion is satisficed; only then may transfers be allocated to the borderline poor. The model’s parameters can represent varying degrees of concern for the poorest, highlighting its flexibility as a framework for welfare evaluation.