{"id":"arxiv/2610.05621","url":"https://arxiv.org/abs/2610.05621","tags":["math.ST","math.PR","stat.ME"],"score":0,"title":"Exact finite-sample bias, variance, and MSE of normalized scale-invariant inequality indices under Gamma mixture models","domain":"math.ST","source":"arxiv","authors":["Roberto Vila","Felipe Quintino"],"reviews":0,"subject":"math.ST","summary":"We study the finite-sample properties of estimators of a broad class of normalized scale-invariant inequality indices (NSIIs) under finite Gamma mixture models. The class includes the Gini coefficient, the $m$th Gini index, the Extended $m$th Gini index, linear order-statistic inequality indices, and several entropy- and dispersion-based measures. Building on the Gamma--Dirichlet representation, we derive exact expressions for the expectation and finite-sample bias of the natural estimator of an NSII based on a $U$-statistic when the population follows a finite Gamma mixture with a common…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-04","reactions":{}}