Empirically Implementing a Social Welfare Inference Framework

This paper builds on recent econometric developments establishing distribution-free statistical inference methods for quantile means and income shares for a sample distribution of microdata to propose an approach to empirically Implement several dominance criteria for comparing economic well-being and general income inequality between distributions. It provides straightforward variance-covariance formulas in a set of practical empirical procedures for formally testing economic well-being and inequality comparisons such as rank dominance, Lorenz dominance and generalized Lorenz dominance between distributions.
The tests and procedures are illustrated with Canadian census data between 2000 and 2020 on women's and men's incomes. It is found that both women's and men's economic well-being statistically significantly improved over this period, while income inequality significantly increased over 2000-15 and then fell over 2015-20.