论文部分内容阅读
本文在一般SV模型的基础上加入泊松跳跃过程,用于描述股价波动过程中存在的暴涨暴跌现象,同时考虑波动率的周内效应。在识别模型中的参数和潜变量时,采用MCMC技术:部分参数采用Gibbs抽样的方法予以解决,对目标函数较为复杂的参数和潜变量采用切片抽样技术。实证分析表明,具有跳跃行为和杠杆效应的SV模型能够更准确地刻画收益率的波动;A股市场在考查期内并不存在明显的杠杆效应,香港股市则有显著的杠杆效应;A股市场在发展的不同时期,都存在显著的星期一效应和星期五效应,表明交易者会利用对政策的预期调整交易策略。
This paper adds the Poisson jump process to the general SV model to describe the ups and downs in the process of stock price volatility, taking into account the weekly effect of volatility. In the identification of parameters and latent variables in the model, the MCMC technique is adopted: some parameters are solved by using Gibbs sampling method, and the slice sampling technique is adopted for parameters and latent variables with more complicated objective functions. Empirical analysis shows that the SV model with jump behavior and leverage effect can more accurately characterize the volatility of returns. There is no obvious leverage effect in the A-share market during the examination period, while the Hong Kong stock market has significant leverage effect. The A-share market There are significant Monday effects and Friday’s effects in different periods of development, indicating that traders will make use of the anticipation of the policy to adjust trading strategies.