A0242
Title: Using quantile time series and historical simulation to forecast financial risk multiple steps ahead
Authors: Richard Gerlach - University of Sydney (Australia) [presenting]
Giuseppe Storti - University of Salerno (Italy)
Antonio Naimoli - University of Salerno (Italy)
Abstract: A method for quantile-based, semi-parametric historical simulation estimators of multiple step ahead value-at-risk (VaR) and expected shortfall (ES) models is extended and developed. The method is based on employing the quantile loss function, analogous to how the quasi-likelihood is employed by standard historical simulation methods. The estimated quantile series is used to scale the returns data, and then re-sampling methods are employed to estimate the forecast distribution one step and multi-steps ahead, allowing tail risk forecasting. The method is extended to allow a measurement equation, thus incorporating realized measures and including realized GARCH and realized CAViaR type models in the class of models it pertains to. The proposed method implicitly assumes and is applicable to any data or model where the relationship between VaR and ES in the conditional return distribution does not change over time; this includes most modern financial time series models. The finite sample properties of this method and its comparison with existing historical simulation methods are evaluated via a simulation study. A forecasting study, applied to 3 indices and 3 assets, assesses the relative accuracy of the 1\% and 2.5\% VaR and ES one-day-ahead and ten-day-ahead forecasting results for the proposed class of models compared to several competitors.
