Sampling Errors and Cross-Country Comparisons of Income Inequality

Abstract

The growing interest in cross-national comparisons of income inequality is primarily a result of the establishment of the Luxembourg Income Study (LIS) database and the wide range of studies on income inequality based on LIS data. The majority of these studies suffer, however, from a major weakness since sampling errors are neither reported nor taken into account when nations are ranked according to estimates of the Gini coefficient or some alternative measure of inequality. This paper discusses the impact of accounting for sampling errors when making comparisons of income inequality across nations.