Quantile forecasts for financial volatilities based on parametric and asymmetric models (Q1726164)

The authors introduce a parametric quantile forecast strategy for interval and Value-at-Risk forecasts. The strategy aims to address asymmetries in three components including the mean, the volatility and the distribution. The authors apply the proposed strategy to the realized volatilities and volatility indices. The main methodology for quantile forecast is discussed in Section 2, where the models for asymmetries in the three components are presented. In Section 3, the asymmetric features are identified using some descriptive statistics from the realized volatility data sets, and the in-the-sample estimation results are provided and discussed. The out-of-sample forecast performance of the proposed method is discussed in Section 4. Specifically, forecast intervals for realized volatilities and volatility indices are studied.