Estimation of Value-at-Risk by a New Model Based on Gaussian Copula and Standardized Standard Asymmetric Exponential Power Distribution Errors for Sovereign Credit Default Swaps

In this paper, we present a novel enhancement to the RiskMetrics methodology initial-ly introduced by J.P. Morgan in 1994. Our approach incorporates a copula-GARCH model combined with an Asymmetric Exponential Power Distribution (AEPD), tai-lored to better address the nuances of Sovereign Credit Default Swaps (CDSs). The core aim of our model is to provide a more accurate analysis of the dependence struc-ture among these financial instruments and to refine the estimation of Value-at-Risk (VaR). We employ three different VaR estimation methods: Historical Simulation, Variance-Covariance, and Monte Carlo simulations. The empirical findings from our study indicate a clear superiority of the SSAEPD-GARCH-Copula model over the tra-ditional RiskMetrics framework, particularly in terms of fitting the model to data and forecasting VaR. Furthermore, through rigorous backtesting, we confirm that the in-troduction of SSAEPD significantly enhances the precision of VaR forecasts. These results substantiate the potential of our modified model as a robust tool in the risk management of Sovereign CDSs, offering substantial improvements over existing methodologies in capturing the complex risk dynamics of these instruments.