基于异质市场假说超高频时间序列建模及应用研究

发布时间：2018-01-06 07:13

本文关键词：基于异质市场假说超高频时间序列建模及应用研究　出处：《长沙理工大学》2012年硕士论文　论文类型：学位论文

【摘要】：股市波动及相关特征是国内外学者研究金融衍生工具的定价、有效资产组合的选择及金融风险管理的关键变量，也是金融学领域的一个研究热点。近年来，随着计算机及通讯技术的快速发展，极大的降低数据记录和存储的成本，从而使得金融高频数据日益成为研究金融资产价格波动及市场微观结构的重要研究对象，同时，也掀起了国内外学者研究金融高频时间序列的热潮。自Engle和Russell（1998）首次提出自回归条件久期模型（Autoregressive Conditional DurationACD）以及Anderson和Bollerslev(1998)首次提出“已实现”波动率的概念以来，基于高频日内数据的建模取得了显著发展。首先，在金融高频时间序列基本特征认识的基础上，本文对金融高频/超高频时间序列的建模问题作了深入探讨，着重对基于高频数据的HAR-RV模型和基于超高频数据的ACD模型从理论推导和参数估计两方面做了详细介绍。其次，，本文从高频时间序列的定义和基本统计特征研究出发，重点从理论上对高频金融数据的基本统计特征进行了归纳和总结，然后利用中国上证综指等时1分钟高频数据对中国股市上的高频时间序列的基本统计性质进行了实证分析，研究发现我国股市高频数据在统计上表现出明显的“尖峰厚尾”的非正态特征，且“日内效应”十分显著。最后，从异质市场假说的角度，基于超高频时间序列构建了HAR-BACD-V模型，并将其应用在中国股市上进行实证分析。实证结果进一步验证了我国股市交易者的异质性，且不同交易频率的投资者的交易行为对我国股市波动的影响是不同的：即短期交易者对股市波动影响最大，中期交易者其次，长期交易者影响最小；另外，还发现微观因子交易量对交易持续期存在明显的负向效应，这从实证的角度进一步验证了Easley和O’Hara (1992)的市场微观结构假说，同时交易量对股市收益率和波动率也有较强的正向影响，这也说明了我国股市的信息传播基本上是符合Copeland（1976）的连续信息到达假说。
[Abstract]:Stock market volatility and its related characteristics are the key variables of financial derivatives pricing, effective portfolio selection and financial risk management, and are also a hot research topic in the field of finance. With the rapid development of computer and communication technology, the cost of data recording and storage is greatly reduced. As a result, high-frequency financial data has become an important research object in the study of financial asset price volatility and market microstructure. Since Engle and Russell 1998), the autoregressive conditional duration model was first put forward (. (Autoregressive Conditional DurationACD) and Anderson and Bollerslev 1998). The concept of "realized" volatility has been proposed for the first time. Modeling based on high frequency day data has made remarkable progress. Firstly, based on the recognition of the basic characteristics of financial high-frequency time series, the modeling of financial high-frequency / ultra-high frequency time series is deeply discussed in this paper. The HAR-RV model based on high frequency data and the ACD model based on UHF data are introduced in detail from two aspects: theoretical derivation and parameter estimation. Secondly, from the definition of high-frequency time series and the study of basic statistical characteristics, this paper focuses on the theoretical analysis of the basic statistical characteristics of high-frequency financial data. Then the statistical properties of the high frequency time series in the Chinese stock market are analyzed empirically by using the high frequency data of 1 minute isochronous time series of the Shanghai Composite Index of China. It is found that the high frequency data of Chinese stock market show obvious non-normal characteristics of "peak and thick tail" statistically, and the "intraday effect" is very significant. Finally, from the perspective of heterogeneous market hypothesis, the HAR-BACD-V model is constructed based on UHF time series. The empirical results further verify the heterogeneity of Chinese stock market traders. The trading behavior of investors with different trading frequencies has different effects on the volatility of China's stock market: the short-term traders have the greatest impact on the stock market volatility, the medium-term traders have the second, and the long-term traders have the least impact; In addition, it is also found that the transaction volume of micro factors has a negative effect on the transaction duration. This further verifies the market microstructure hypothesis of Easley and Ogan Hara 1992 from the empirical point of view, and the trading volume also has a strong positive effect on the stock market return and volatility. This also shows that the information transmission in China's stock market is basically consistent with the hypothesis of continuous information arrival proposed by Copelandin1976.
【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】：F832.51;F224

【参考文献】