Peaks over thresholds modelling with multivariate generalized Pareto distributions
Preprint, 2017

The multivariate generalized Pareto distribution arises as the limit of a normal- ized vector conditioned upon at least one component of that vector being extreme. Statistical modelling using multivariate generalized Pareto distributions constitutes the multivariate analogue of univariate peaks over thresholds modelling. We exhibit a construction device which allows us to develop a variety of new and existing para- metric tail dependence models. A censored likelihood procedure is proposed to make inference on these models, together with a threshold selection procedure and several goodness-of-fit diagnostics. The models are fitted to returns of four UK-based banks and to rainfall data in the context of landslide risk estimation.

financial risk measurment

extreme values



Anna Kiriliouk

Holger Rootzen

Chalmers, Matematiska vetenskaper, matematisk statistik

Göteborgs universitet

Johan Segers

Jennifer L. Wadsworth



Sannolikhetsteori och statistik


Building Futures



Grundläggande vetenskaper