The EWMA approach based on two different distributions and Value at Risk-Evidence from Shanghai and Shenzhen Stock Exchange
The Exponentially weighted moving average (EWMA) approach is commonly used to estimate conditional volatility of short horizon asset returns; it is based on the maximum likelihood estimator of variance of the normal distribution. While there are much evidence shows that the conditional distribution is leptokurtic and fat tail rather than normal distribution. Guermat and Harris (2000) recommended that in estimating the Value-at-Risk the EWMA approach based on Laplace distribution were always more robust than the one based on normal distribution. We employ the methodology similar to them and made empirical test on Chinese stock market. Moreover, we use GARCH model to find the optimal decay factor for each series. Our conclusion is that there is no key difference between the two models and both of them can accurately and efficiently in forecasting the VaR, and for some particular case certain model is more suitable.
Value-at-risk exponentially weighted moving average (EWMA) Laplace distribution GARCH model
Li SHEN Xuedong ZHENG
Department of Finance,School of Mathematics and Statistics,School of Finance,Zhejiang University of Finance & Economics,Hangzhou 310018
国际会议
2008 International Conference on Risk and Relianility Management(2008风险与可靠性管理国际会议)
北京
英文
55-59
2008-11-10（万方平台首次上网日期，不代表论文的发表时间）