Do the intradeciles inequalities matter? A note

Using an algorithm suggested by Berrebi and Silber (1987) and assuming that the population within any decile is distributed uniformly between its boundsa andb, the paper shows that the intradecile inequality is generally not important and may be ignored. This finding is useful in empirical analysis, even today, when inequality indices are sought. Although current microdata sets from which the Gini indices can be calculated directly are widely available, past data are still in use and the new data are not always accessible to the interested user in micro detail, while the deciles bounds are very frequently reported. The study also points to a new constant. The Gini of a variable over the interval [0,b] when there is a large number (n>100) of individuals and the variable is distributed uniformly, is 1/3 regardless of the value ofb.