An inequality decomposition method which minimizes equivalence scales contamination problems

Decomposable measures are a useful tool to analyze the impact of households characteristics on income or expenditure inequality. However, the results are sensitive to the choice of equivalence scales in a heterogenous population. In this paper, we assume that equivalence scales depend only on the number of persons in the household. In this context, we suggest a method to free the decomposition analysis from the possible 'contamination' that will arise if we use an inappropiate equivalence scale. The method is applied to the evolution of the standard of living in Spain during the 80' s. We study the structure of Spanish inequality in 1980-81 and 1990-91, as well as the trend in overall inequality over time in terms of three factors: i) the change in within-group inequality, ii) the change in between-group inequality, and iii) the demographic change across partition subgroups