Review on Three New Value at Risk (VaR) Models

Abstract

The emergence of financial derivatives complicates traditional financial products and increases financial market volatility. Individuals and financial institutions are both exposed to more complex and uncontrollable risks in this environment. Because of the risk's uncertainty, we must use reasonable methods to predict and estimate it in order to achieve the goal of risk control. This paper discusses three new VaR (Value at Risk) models that have emerged in recent years based on the ARCH family model using a method of literature review. The ARMA-EGARCH model, for example, combines the ARMA model to describe constant variance time series and the EGARCH model to describe heteroscedasticity phenomena, and theoretically can better describe the fluctuations of financial time series and obtain an independent time series with the same distribution. The sequence is processed using extreme value theory, which is the ARMA-EGARCH-GPPD model, in conjunction with the GPD model. We used the ARMA-EGARCH-semi-parametric method in conjunction with the historical simulation method and the parameter method to avoid cumbersome quantile calculation because the model algorithm is more complex. The generalized EWMA risk value prediction model has more advantages for financial data with large peaks.