Forecasting VaR and ES Using A Factorial Hidden Markov Approach
VaR and ES are two important measures that estimate the risk of financial assets.They are of great significance as they can help financial risk management and financial crisis identification.In this paper,a latent variable,constructed by the Factorial Hidden Markov Model,is added into the CAViaR model,and a new model to jointly predict VaR and ES-FHM-CAViaR Model is finally created.In this model,the latent variable is driven by a Factorial Hidden Markov Model to reflect the long-term and short-term impact of market and the Markov chain can transform among hundreds of states.Then,we use this proposed new model to analyze logarithmic return data of different stock markets.Results show that FHM-CAViaR Model has achieved superior prediction effects and the aggregation of VaR increases significantly during the crisis.In general,the model proposed in this paper can recognize the regime switching of the market through the latent variable and estimate the VaR and ES of financial assets more accurately,which provides a new approach to measuring financial risk.
万方数据
维普
