Department seminar by Andreas Tsanakas

Please Note: This seminar will be given online.

Multivariate stress testing by change of measure

For decision makers whose portfolios are exposed to functions of a random vector of risk factors, we consider a stress to be a change of measure, placing a higher weight on scenarios of interest. In particular, we define a stressing mechanism as a mapping from risk factors to Radon-Nikodym densities. We focus on stressing mechanisms that are invariant to monotonic transformations of risk factors and increase their joint tail probabilities. We study the properties of two families of stressing mechanisms, based respectively on mixtures of univariate stresses and on transformations of statistics we call Spearman and Kendall's cores. Furthermore, the specified stressing mechanisms can be used to derive new capital allocation methods, with properties different to those typically found in the literature.  The proposed methods are applied to the simulation model of a UK-based non-life insurer.