New model revolutionizes financial forecasting, improves risk estimation accuracy.
The article explores using quantile methods to improve forecasting of asset returns. By incorporating long memory components and past return information, the model provides better estimates and forecasts for Value at Risk (VaR). The researchers also enhance the model by considering the dynamics of scale and shape separately, leading to more accurate VaR forecasts. By extending the model to include multiple probability levels and dynamic scale, they are able to explain time-varying patterns between quantiles. Bayesian inference is used to address estimation issues and provide more robust estimates. Overall, the study aims to improve the understanding of conditional asset return distributions and enhance forecasting accuracy for financial applications.