The Quantile Garch-Distributed Lag (QGDL) framework: a unified model for high-volatility time series

Abstract

The essence of this article is that it has developed the Quantile GARCHDistributed Lag (QGDL) model as a recent and superior model to the QARDL model. Although the QARDL model is only able to deal with asymmetric relationships in the quartiles and the assumption of homoscedasticity, our model is able to address this weakness using the simultaneous volatility clustering. The model that we propose is the most effective and correct in estimating the parameters in cases of changing conditional variance with time, which makes it the most appropriate measure to use in the analysis of extremely volatile financial time series.