Income-expenditure surveys typically provide incomes on the household level. As households can differ in size and needs, a reliable assessment of inequality in living standards, therefore, necessitates the conversion of the original heterogeneous into an artificial quasi-homogeneous population. Ebert and Moyes (2003) and Shorrocks (2004) theoretically explore the properties of two alternative conversion strategies: a weighting of household equivalent incomes by size and by needs. We use data from the Luxembourg Income Study for examining the sensitivity of the Gini and the Theil index to the chosen conversion strategy, and explain our results by means of an inequality decomposition by population subgroups.
