Extreme incomes and the estimation of poverty and inequality indicators from EU-SILC

Micro-data estimates of welfare indices are known to be sensitive to observations from the tails of the income distribution. It is therefore customary to make adjustments to extreme data before estimating inequality and poverty statistics. This paper systematically evaluates the impact of such adjustments on indicators estimated from the EU-SILC (Community Statistics on Income and Living conditions) which is expected to become the reference source for comparative statistics on income distribution and social exclusion in the EU. Emphasis is put on the robustness of cross-country comparisons to alternative adjustments. Results from a sensitivity analysis considering both simple, classical adjustments and a more sophisticated approach based on modelling parametrically the tails of the income distribution are reported. Reassuringly, ordinal comparisons of countries are found to be robust to variants of data adjustment procedures. However, data adjustments are far from innocuous. Cardinal comparisons of countries reveal sensitive to the treatment of extreme incomes, even for seemingly small adjustments.