Stochastic dominance for super heavy-tailed random variables

Abstract

We introduce a class of super heavy-tailed distributions and establish the inequality that any weighted average of independent and identically distributed super heavy-tailed random variables stochastically dominates one such random variable. We show that many commonly used extremely heavy-tailed (i.e., infinite-mean) distributions, such as the Pareto, Fr\'echet, and Burr distributions, belong to the class of super heavy-tailed distributions. The established stochastic dominance relation is further generalized to allow negatively dependent or non-identically distributed random variables. In particular, the weighted average of non-identically distributed random variables stochastically dominates their distribution mixtures. Applications of these results in portfolio diversification, goods bundling, and inventory management are discussed. Remarkably, in the presence of super heavy-tailedness, the results that hold for finite-mean models in these applications are flipped.