Rekonstrukcja światowego rozkładu dochodów na podstawie minimalnej informacji statystycznej

The aim of this paper is to obtain a sample from the world income distribution (WID). We assumed log-logistic form of countries’ income distributions with unit means. This implies that such distributions are fully described by their corresponding Gini indices. Gini estimates came from Deninger-Squire, WID2, and other databases for 119 countries in the years 1990-2005. We generated random sample for every distribution and multiplied its values by GDP/capita. Sample size was controlled by sequential ratio test. The world sample consisted of country samples weighted by population shares. We have found that WID is bimodal with diminishing inequality and poverty during analyzed period.