GARCH模型族在沪深300中的比较研究

风险测量一直是金融研究领域的热门话题,而如何构建合适的模型来衡量风险自然而然成为众多学者研究的关注点.VaR方法是当今应用最广泛的衡量金融风险的方法之一,其核心又在构建良好的波动率估计模型.GARCH模型族能很好地描述股指波动率呈现的重尾、波动性聚集、杠杆效用等,是当前效果比较好的条件异方差性的模型.本文着重研究基于GARCH模型族(GARCH、EGARCH、PGARCH)在不同分布假定下(高斯分布、t分布、广义误差分布)的表现,从而计算出沪深300的在险价值( VaR),比较分析模型拟合效果,选出适合的模型,对规范国内沪深300的风险管理提供了理论依据.

A Comparative Study on the GARCH Models in CSI 300

Risk measurement has always been a hot topic in the realm of financial research,and how to build an appropriate model to assess the risk has become the focus of many scholars’ research.VaR(Value at Risk) is one of the most widely used methods to measure financial risk nowadays,whose core is to build a good volatility estimation model.GARCH Models are considered good conditional heteroscedasticity models at the moment given its well capability of describing heavy tails,volatility clustering and leverage effects of the index’s volatility.This paper delves into the fitting results of GARCH Models(GARCH,EGARCH and PGARCH) under different distribution assumptions (Gaussian distribution,t distribution and generalized error distribution),and hence calculates the VaR of CSI 300, compares and analyzes the fitting performance among the models and picks the most appropriate model as the theoretical basis for standardizing risk management in CSI 300.