A Sequence of Nested Exponential Random Variables with Connections to Two Constants of Euler

Abstract: We investigate a recursively generated sequence of random variables that begins with an Exponential random variable with parameter (i.e., inverse-mean) 1, and continues with additional Exponentials, each of whose random parameter possesses the distribution of the prior term in the sequence. Although such sequences enjoy some (limited) applicability as models of parameter uncertainty, our present interest is primarily theoretical. Specifically, we observe that the implied sequence of distribution functions manifests a surprising connection to two well-known mathematical quantities first studied by Leonhard Euler: the Euler-Gompertz and Euler-Mascheroni constants. Through a close analysis of one member of this distribution-function sequence, we are able to shed new light on these constants.