B0471
Title: Angular volatility for multivariate extremes
Authors: Miguel de Carvalho - FCiencias.ID - Associacao para a Investigacao e Desenvolvimento de Ciencias (Portugal) [presenting]
Simone Padoan - Bocconi University (Italy)
Abstract: The angular volatility is proposed as a measure of the dynamics of extremal dependence over time. The building block for the notion of angular volatility is that of a family of time-changing spectral densities of a multivariate extreme value distribution. For modeling, our preferred implementation is through a stochastic volatility method which we fit through Bayesian methods.