基于贝叶斯原理的随机波动率模型分析及其应用  

Stochastic Volatility Models Based Bayesian Method and Their Application

作　　者：蒋祥林[1] 王春峰[2] 

机构地区：[1]复旦大学金融研究院,上海200433 [2]天津大学金融工程研究中心,天津300072

出　　处：《系统工程》2005年第10期22-28,共7页Systems Engineering

基　　金：国家杰出青年科学基金资助项目(70225002);上海市哲学社会科学规划资助项目(2005BJB002)

摘　　要：基于贝叶斯原理,对随机波动性模型进行研究,并将随机波动率模型应用股市风险价值V aR的估计与预测。针对中国股市数据进行的实证结果表明,与GARCH模型相比,随机波动率模型能更好地描述股票市场回报的异方差和波动率的序列相关性;基于随机波动率的V aR较GARCH模型的V aR具有更高的精度。Stochastic volatility （SV） models here are investigated based on bayesian method, and are applied to estimate and forcast the Value at Risk （VaR） of the Chinese stock market. Empirical results on Chinese stock market indicate that stochastic volatility model outperforms the ARCH model in capturing the heteroskedasticity and serial correlation of volatility of the stock market. VaR based on SV models is more precision than that based on GARCH models.

关 键 词：随机波动率模型 GARCH模型 风险价值 贝叶斯原理 

分 类 号：F830.91[经济管理—金融学]