连续时间SV模型的估计及在金融风险分析中的应用  

The Estimation of Continuous Time Stochastic Volatility (SV) model and its application in the analysis of financial risk

作　　者：孟利锋[1] 张世英[2] 

机构地区：[1]天津商业大学商学院,天津300134 [2]天津大学管理学院,天津300072

出　　处：《金融理论与实践》2008年第9期10-13,共4页Financial Theory and Practice

基　　金：国家自然科学基金资助项目(70171001)

摘　　要：本文分析了由连续时间平方根随机波动模型确定的资产收益的无条件联合特征函数及统计特性。对连续时间SV模型,提出了基于经验特征函数的参数估计方法,这种估计方法既不需要对连续时间过程的离散化,也不需要取样路径的模拟,实施起来较简单。使用上海和深圳股市的指数日收益数据对经验特征函数方法在连续时间随机波动模型参数估计中的应用进行了实证分析,并证明了两个市场波动过程存在均值回复及连续时间随机波动模型拟合资产收益数据的优势。This article analytically obtained the unconditional joint characteristic function and statistic characters of asset return which determined by continuous time square-root stochastic volatility （SV） model. Parameter estimation method based on empirical characteristic function （ECF） was proposed to estimate continuous time SV model. This simple method requires neither discretization nor simulation. The estimation of continuous time square-root SV model using ECF method was illustrated, along with an empirical application using Shanghai and Shenzhen stock index data. The results suggest that there exist mean reversion in the volatility process of these two markets, and the advantages of fitting asset return data using continuous time SV model.

关 键 词：连续时间SV模型 经验特征函数 参数估计 杠杆效应 

分 类 号：F831.59[经济管理—金融学]