NOTE ON THE MOMENT GENERATING FUNCTION OF THE MULTIVARIATE NORMAL DISTRIBUTION

Abstract

We present a streamlined proof of a formula for the derivatives of the moment generating function of the multivariate normal distribution. We formulate it in terms of the summation of the contractions by pairings, which encodes a combinatorial computation procedure. We give two applications. First, we provide a simple proof of Isserlis’ theorem and derive a formula for the moments of the multivariate normal distribution.Second, we calculate the moments of the product of a finite number of correlated normally and lognormally distributed random variables.