International Comparison of Household Inequalities : Based on Micro Data with Decompositions

This paper demonstrates the usefulness of the decomposability property of the Generalized Entropy (GE) family of measures in comparing inequality among countries. A family of Generalized Entropy measures are decomposed by family size and by the household head's age, gender, education, and ethnicity. This is done in order to learn about components which are due to demographic differences "between" households, and "within" group components which are free of such group characteristics. This will further our understanding of the impact of different social-economic structures upon the distribution of income. Looking at the overall inequality for comparative analysis without the decompositions can provide us with only a partial picture of the differences and thus is inadequate. Moreover, internal analysis is enhanced since the decompositions will locate the potential source of inequality for diagnostic policy purposes. Luxembourg Income Study data sets are chosen for their richness and comparability of micro data on variables and attributes such as income, age, education, family size, gender, and ethnicity