This talk has been automatically captioned. You can remove these by pressing CC on the video toolbar.

Name: Karla Vianey Ramirez

Talk Title: Bayesian semiparametric modeling of jointly heteroscedastic extremes

Abstract: We introduce a Bayesian semiparametric model for learning about the magnitude and frequency of joint extreme values. The joint scedasis function for joint extremes is here devised as a function that carries information on the frequency of joint extremes over time. We develop Bayesian estimators for the two parameters in the model—the joint scedasis function and the coefficient of tail dependence; to learn about the joint scedasis function we resort to finite mixtures of Polya trees, as they can be used to define a flexible prior in the space of scedasis functions. The simulation study shows that the proposed methods are able to recover the true magnitude and frequency of joint extremes in a variety of simulation scenarios. An application of the proposed methodology to the so-called FAANG (Facebook, Apple, Amazon, Netflix and Alphabet’s Google) stocks reveals some interesting insights on the dynamics governing their joint extreme losses over time.

This talk is a contributed talk at EVA 2021.