Inequality and welfare estimates using two alternative weighting schemes

Income and expenditure surveys typically provide data on the household level, and household types may differ in needs. In order to make the standard measures of inequality and welfare applicable to such a heterogeneous population, researchers transform it into an artificial quasi-homogeneous population. The transformation requires besides the implementation of an 'appropriate' income concept the choice of an 'appropriate' weight assigned to each income receiving unit. Ebert and Moyes (2003) and Shorrocks (2004) suggest two alternative weighting schemes that satisfy different sets of principles. Using representative household-level income data for 20 European countries, we study the quantitative impacts of the two weighting schemes on two central measures, the Gini coefficient and mean equivalent income, respectively.