INCOME DISTRIBUTION MODELS AND INCOME INEQUALITY MEASURES FROM THE ROBUST STATISTICS PERSPECTIVE REVISITED

Considerations related to income distribution and income inequalities in populations of economic agents belong to the core of the modern economic theory. They appear also in a public debate concerning postulates as to taxation or pension politics, in theories of a human capital creation or searching for regional development factors. Results of statistical inference conducted for giving arguments pro or against particular hypotheses, strongly depend on properties of statistical procedures used within this process. We mean here for example: a quality of probability density estimator in case of missing data, a quality of skewness measure in multivariate case departing from normality, or a quality of dimension reduction algorithm in case of existence of outliers. In this paper from the robust statistics point of view, we analyse difficulties related to statistical inference on income distribution models and income inequalities measures. Theoretical considerations are illustrated using real data obtained from Eurostat and Minessota Population Center (IMPUS)