摘要
高频金融时间序列的分析与建模是金融计量学的一个崭新的研究领域,已实现波动和已实现极差波动是针对高频金融时间序列而开发的两种全新的波动率度量方法.首先证明了在理想状态下,已实现极差波动比已实现波动是更有效的波动估计量,然后基于渐近关系讨论了高频数据最优抽样频率问题.在模拟试验的基础上,比较了微观结构效应对两种波动率度量方法的影响程度.最后,通过实证分析对上证综指的高频数据给出一个最优抽样频率.
High-frequency financial time series analysis and modeling is a new research field in financial econometrics, and realized range-based volatility and realized volatility are new measure approaches of volatility in high-frequency data field. At first, the paper proves that the realized range-based volatility is more efficient than the realized volatility in estimating volatility in ideal situation. Then, based on asymptotic relationships, the optimal sampling of high-frequency financial data is discussed. Under simulated tests, the effects of microstructure on realized range-based volatility and realized volatility are compared. Lastly, through empirical analysis of composite index of Shanghai Stock Market, the optimal sampling of high-frequency data is given.
出处
《系统工程学报》
CSCD
北大核心
2007年第4期437-442,共6页
Journal of Systems Engineering
基金
国家自然科学基金资助项目(70471050)
关键词
高频金融数据
已实现极差波动
最优频率
微观结构
high-frequency financial data
realized range-based volatility
optimal sampling
microstructure
