摘要
在实证金融研究中,波动持续性是一个非常重要的研究问题。资产定价理论表明,在随机波动建模中,波动持续性可以通过检验单位根来反映。然而,对于随机波动模型来说,经典的单位根检验统计量如ADF等检验统计量应用非常困难。本文则在贝叶斯框架下,针对具有协变量的厚尾分布金融随机波动模型,给出了检验单位根的方法。蒙特卡罗模拟结果显示本文给出的方法能够取得非常好的检验功效。最后,我们用万科地产的周收益率数据论证本文给出的方法.
In empirical research, volatility persistency is a very important topic. Asset pricing theory shows that persistency can be reflected by testing unit root in stochastic volatility models. However, classical unit root testing statistics such as ADF can not be applied easily to test unit root for stochastic volatility models with heavy tails and covariate variables. This paper aims to provide a Bayesian unit root testing approach for these models. Monte Carlo simulation studies show that this new test statistic can obtain good power. Finally. this approach is illustrated by employing the return series of Wanke Real Estate Company
出处
《中大管理研究》
2011年第4期100-111,共12页
China Management Studies
关键词
单位根
贝叶斯因子
厚尾随机波动模型
贝叶斯检验
波动持续性
unit root test, bayes factor, financial stochastic volatility models with heavy taildistribution, bayesian test, volatility persistency