DPOT methodology: An application to value-at-risk

Abstract

Threshold methods, based on fitting a stochastic model to the excesses over a threshold, were developed under the acronym POT (peaks over threshold). To eliminate the tendency to clustering of violations, a model-based approach within the POT framework, which uses the durations between excesses as covariates, is presented. Based on this approach we suggest models to forecast one-day-ahead Value-at-Risk and apply these models to the Standard & Poor’s 500 Index. Out of sample results provide evidence that they can perform better than state-of-the art risk models.