Gini index (G), a scaled version of Gini’s mean difference (GMD), is a U-statistic of degree two. In this presentation, I will present a general approach to construct a new class of inequality measures called ad-hoc inequality measures (AIM) which are based on U-statistics of degree higher than two. The new AIMs satisfy anonymity, scale invariance, and population independence.We illustrate situations where delicate internal features with income disparities are more clearly explained and motivated from elementary economic persuasions of population dynamics, but those delicate features may be incorrectly missed by G. That is, one or more newly proposed AIMs are more apt to capture intricatefeatures than G in some instances. Without assuming any specific nature of the population distribution for the data, we have derived (i) the asymptotic mean square error of a general AIM; and also (ii) the asymptotic distribution of AIM: These results have provided useful inference methodologies which have beensupplemented with extensive sets of simulations and analyses of real economic data.
Venue
SMS seminar Hall
Speaker
Bhargab Chattopadhyay
Affiliation
IIIT Vadodara
Title
Generalized Gini Index based on U-statistics