论文部分内容阅读
反映金融市场条件波动时变性的模型都只是对资产真实数据生成过程的一种近似拟合,许多模型严格来说存在设定偏误,这样可能导致由它们所预测的条件波动相对真实值存在偏差。本文用一回归模型检验一般正态VaR估计中的简单平均方法估计的条件波动是否存在偏差,另一方面用它对这一条件波动估计值进行矫正,并以此为基础估计相应的VaR。通过对上证指数和深成指数1995-2005年日收益率数据的分析,我们发现,用简单平均方法对我国股市波动进行估计存在非常大的条件偏差,并发现波动矫偏后所预测的VaR值的准确性和有效性都明显提高。
A model that reflects the time-variability of financial market conditions is only an approximation to the real-life data generation process for many assets, with many models strictly speaking set-errors that may cause deviations from the true values of the conditions they predict . In this paper, a regression model is used to test the existence of deviation of the conditional volatility estimated by the simple average method in the normal normal VaR estimation. On the other hand, it is used to correct the conditional fluctuation estimate and estimate the corresponding VaR based on this. Through the analysis of the daily rate of return of the Shanghai Composite Index and the Shenzhen Composite Index from 1995 to 2005, we find that there is a very large conditional deviation in the estimation of the fluctuation of the Chinese stock market by the simple average method, and the VaR value predicted by the volatility correction The accuracy and effectiveness are significantly improved.