A change-point approach for the identification of financial extreme regimes

Inference over tails is usually performed by fitting an appropriate limiting distribution observations that exceed a fixed threshold. However, the choice of such threshold critical and can affect inferential results. Extreme value mixture models have been defined to estimate using full dataset give accurate tail estimates. Such assume behavior constant for all observations. extreme financial returns often changes considerably in time occur sudden shocks market. Here model class extended formally take into account distributional change-points, allowing presence regime-dependent parameters modelling distribution. This extension uses both thresholds change-point locations, giving uncertainty measures quantities. Estimation functions interest analyses via MCMC algorithms. Our approach evaluated through series simulations, applied real data sets assessed against competing approaches. Evidence demonstrates inclusion different regimes outperforms static dynamic approaches applications.