No abstract

In this paper a method for analysing the fairness of an income tax system when portioning the population into heterogeneous socio-economic groups is proposed. The equitable tax system is defined by the three axioms given by Kakwani and Lambert (1998) and, as they suggest, inequity is evaluated by the negative influences on the redistributive effect of the tax associated with axiom violations. Measuring the extent of axiom violations among households belonging to different groups, we improve the Kakwani and Lambert analysis, which is able to detect only the existence of overall inequities. We propose a method that allows for evaluation of the contribution of each group to the overall inequity. Moreover, the adopted method enables disentangling the directions of violations. The obtained results allow us to judge how axiom violations discriminate among groups in their reciprocal relationships. An application to the 2010 Italian income tax reveals that inequities disproportionately penalize the household typologies. More precisely, unfairness affects households with children more severely than the other household groups. JEL Codes: C81, H23, H24

Generally, inequality indices play a basic role in the analysis of welfare economics, also appearing as technical tools applied to income data. A good deal of findings in this research field is provided by the Gini coefficient, typically used for non-negative income values. Even if negative income is often an unfamiliar concept, its presence in real surveys may lead to difficulty in applying the classical Gini-based inequality measures, as they lie outside their standard ranges. In this paper, the more general issue of negative values is considered and a reformulation of the main Gini-based inequality measures adjusted for the problem of negative values is adopted with the purpose of providing theoretical extensions for the income decomposition approach by both income sources and area components. Investigations about the related inferential issues, conducted thorough simulation studies based on resampling techniques, highlight how the traditional approach of removing negative income values may yield different results in terms of inequality estimation, proving that the proposed approach, based on preserving negative values, is the more appropriate practice to follow to avoid the loss of data that really provide a coherent picture of the inequality condition.

Kakwani and Lambert (1998) state three axioms which should be respected by an equitable tax system. Using the Atkinson-Plotnick-Kakwani re-ranking indexes of taxes, tax rates and post-tax incomes, calculated with respect to the ranking of pre-tax income distribution, they then propose a measurement system to evaluate the negative influences that axiom violations exert on the redistributive effect of taxes. In this paper we reconsider the way Kakwani and Lambert measure violations of their second axiom, which concerns the re-ranking of tax rates. We construct a non-negative index which is strictly faithful to Kakwani and Lambert's commands; we show that the Authors' measure does not exactly fit the statements made in their second axiom. Both Kakwani and Lambert's original measurement system and the modified one are then applied to Italian personal income tax in 2008. According to the modified measurement system, the average tax rate seems to play a smaller role than that suggested by the results gained using Kakwani and Lambert's original methodology.