Dynamic optimality in optimal variance stopping problems

In an optimal variance stopping (O.V.S.) problem one seeks to determine the stopping time that maximizes the variance of an observed process. As originally shown by Pedersen (2011), the variance criterion leads to optimal stopping boundaries that depend explicitly on the initial point of the process. Then, following the lines of Pedersen and Peskir (2016), we introduce the concept of dynamic optimality for an O.V.S. problem, a type of optimality that disregards the starting point of the process. We examine when an O.V.S. problem admits a dynamically optimal stopping time and we illustrate our findings through several examples.