Lorenz curves and second-order dominance criteria are known to be sensitive to data contamination in the right tail of the distribution. We propose two ways of dealing with the problem: (1) Estimate Lorenz curves using parametric models for income distributions, and (2) Combine empirical...

We decompose the generalized Lorenz order into a size and a distribution component. The former is represented by stochastic dominance, the latter by the standard Lorenz order. We show that it is always possible, given generalized Lorenz dominance between two distributions F and G, to find...

This paper estimates a variety of inequality indices to study the evolution of income and consumption inequality in Poland between 1998 and 2008. We use robust methods to adjust for the impact of extremely large observations. We also conduct statistical tests on inequality changes using methods,...

Research background: studies of structures of incomes distributions have been performer for about 15 years. They indicate that there is no one model which describes the distributions in their whole range. This effect is explained by the existence of different mechanisms yielding to low-medium...

Assessments of social welfare do not usually take into account population sizes. This can lead to serious social evaluation flaws, particularly in contexts in which policies can affect demographic growth. We develop in this paper a little-known though ethically attractive approach to correcting...

We decompose the generalized Lorenz order into a size and a distribution component. The former is represented by stochastic dominance, the latter by the standard Lorenz order. We show that it is always possible, given generalized Lorenz dominance between two distributions F and G, to find...

This paper studies the local robustness of estimators and tests for the conditional location and scale parameters in a strictly stationary time series model. We first derive optimal bounded-influence estimators for such settings under a conditionally Gaussian reference model. Based on these...