基于加权已实现极差的中国股市波动特征研究

发布时间：2018-01-01 19:35

本文关键词：基于加权已实现极差的中国股市波动特征研究　出处：《长沙理工大学》2012年硕士论文　论文类型：学位论文

【摘要】：随着计算机及通信技术的发展，当获取金融高频数据成为可能后，，如何运用高频数据进行建模并估计波动率成为当今研究的热点问题之一。已实现极差波动是针对高频时间序列而提出的一种全新的波动率度量方法，而加权已实现极差可以有效的去除波动的“日内效应”，是比已实现极差更有效的波动估计量。一般而言，不同的股票市场表现出不同的波动特征，波动率估计量的优劣性在于能否精确地刻画出股票市场波动的典型特征及变化趋势。大量研究表明，中国股票市场的波动呈现出尖峰厚尾，自相关性和非对称等特征，并且容易受交易量、价差等市场微观因子的影响，基于中国股市高频数据构造出的加权已实现极差是否也能够刻画出上述的波动特征，目前还没有相关文献对此进行研究。本文首先基于沪深300股指的五分钟高频数据构造已实现波动、已实现极差和加权已实现极差序列，通过理论和实证比较分析，已实现极差序列的方差是已实现波动序列方差的五分之一，加权已实现极差可以有效的去处“日内效应”，且具有更稳定的序列特征，证明了基于中国股市高频数据加权已实现极差为更有效的波动率估计量。随后，为了全面考察加权已实现极差波动的特征和预测未来波动，本文对加权已实现极差进行建模，通过分析其序列性质，本文运用自回归模型（AR模型）对其对数序列进行建模，并在该模型的基础上分别添加非对称变量及交易量因素，研究加权已实现极差序列的非对称特征以及交易量对其的影响。实证结果表明：中国股票市场上加权已实现极差序列具有尖峰厚尾、集聚性、持续性等特征，在模型中加入非对称及交易量因素后，模型的预测能力增强，表明加权已实现极差具有非对称性特征，并且与交易量存在较强的正相关关系。
[Abstract]:With the development of computer and communication technology, when the acquisition of financial high frequency data as possible, how to use high frequency data modeling and estimation of volatility has become one of the hot issues of current research. The realized range based volatility is due to high frequency time series and proposed a new method to measure the volatility, and the weighted realized range can be effective remove the fluctuation of intra day effect, is more effective than the realized range based volatility estimators.
Generally speaking, different stock markets show different volatility features, volatility estimates of the amount of quality is the ability to accurately depict the typical characteristics and trend of fluctuations in the stock market. A large number of studies show that China the volatility of the stock market showing a peak thick tail, autocorrelation and asymmetric features, and is easily affected by the transaction the amount of price impact, micro market factor, weighted China stock market high frequency data structure of the realized range is also can be used to describe the volatility characteristics based on the above, there is no relevant literature to study.
Firstly, based on the CSI 300 index five minute high-frequency data structure realized volatility, realized range and weighted realized range sequence, through theoretical and empirical analysis, realized variance range sequence is realized volatility variance weighted 1/5, have been achieved range can effectively place intra day effect, and with the sequence characteristics of more stable, proved that the high frequency data China stock market based on weighted realized range rate estimation for more effective volatility.
Then, in order to fully investigate the weighted realized range based volatility characteristics and predict the future volatility, the weighted realized range based modeling, through the analysis of the sequence properties, using the autoregressive model (AR model) to model the logarithmic sequence, and add respectively factors asymmetric variables and trading volume on the basis of the model study, the weighted realized effect of asymmetric characteristics of poor sequence and the trading volume. The empirical results show that: China stock market weighted realized range based sequence has peak thick tail, agglomeration, characteristics of persistent, asymmetric add factors and trading volume in the model, to enhance the predictive ability of the models show that the weighted realized range is asymmetrical, and there is a strong correlation with trading volume.

【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51

【参考文献】