An identity for expectations and characteristic function of matrix variate skew-normal distribution with applications to associated stochastic orderings (Q6169187)

Stochastic orders, which are partial orders on a set of random variables, provide methods to describe intuitively random variables being larger, riskier or more dependent. They have found applications in actuarial science, economics, comparison of experiments, reliability analysis and queueing theory. In this paper, the authors first establish an identity for $Ef(Y)-Ef(X)$, when $X$ and $Y$ both have matrix variate skew-normal distributions and the function f satisfies some weak conditions. They then derive the characteristic function of the matrix variate skew-normal distribution and they use it to derive some necessary and sufficient conditions for the comparison of matrix variate skew-normal distributions under usual stochastic order, convex order, increasing convex order, upper orthant order, directionally convex order and super modular order. It will be of interest to generalize the present results to matrix variate skew elliptical family of distributions, and the authors state that they are currently working on this problem.