基于GARCH模型高频数据极值的波动性研究

发布时间：2018-08-26 06:57

【摘要】：波动性是衡量金融市场质量和效益的指标,若能精确地捕捉到金融时间序列的波动性特征,就可以极大的提高金融市场的流动性和资源配置。在经济全球化的今天,信息是连续影响证券市场价格的运动过程。数据的离散采集直接影响到获取市场信息的多少,数据采集频率越高,获取的市场信息也就越多。高频数据是对金融市场感兴趣的人们的主要研究对象,尤其是交易人员通过观察高频或逐点数据来确定交易决策。高频数据的极值序列是不同频率的数据在每一段时间间隔内取极大值和极小值而形成的两个时间序列,它与以往研究的时间序列有所不同,能够使我们更准确的分析极端市场的条件和信息对证券市场产生的影响,这也是本文的创新点。研究高频数据以及收益率的统计特征,并刻画市场信息与极值及其收益率之间的波动性,对我国的证券市场进行理论和实证分析,具有非常重要的理论和实际意义。本文主要讨论了基于GARCH模型高频数据极值的波动性研究,选取中国股市中的沪深300指数15分钟收益数据的极大值和极小值为研究对象,运用GARCH模型建模分析。第一章主要是介绍了国内外关于波动性的研究现状。第二章系统的介绍了波动性计量模型的基本理论和波动性的主要特征,并从理论上进行各类模型间的比较,以初步揭示各个模型的不同特点。第三章主要运用GARCH模型对收益序列进行了建模,对波动性的波动聚集性和尖峰厚尾性及波动的非对称性进行了分析。结论表明,沪深股市具有明显的波动性,收益率数据自身存在尖峰厚尾性,波动聚集性,且服从非正态分布。最后总结本文的分析结果,对模型提出了改进。
[Abstract]:Volatility is an index to measure the quality and benefit of financial market. If we can accurately capture the volatility characteristics of financial time series, we can greatly improve the liquidity and resource allocation of financial market. In the economic globalization today, the information is the movement process which affects the stock market price continuously. Discrete data acquisition directly affects the amount of market information, the higher the frequency of data acquisition, the more market information will be obtained. High frequency data is the main research object of people who are interested in financial market, especially traders determine trading decision by observing high frequency or point by point data. The extreme value sequence of high frequency data is two time series formed by taking maximum value and minimum value of different frequency data in each time interval, which is different from previous time series. It can make us more accurate analysis of the extreme market conditions and the impact of information on the securities market, which is also the innovation of this paper. It is of great theoretical and practical significance to study the statistical characteristics of the high frequency data and the return rate and to describe the volatility between the market information and the extreme value and the rate of return. The theoretical and empirical analysis of the securities market in China is of great theoretical and practical significance. This paper mainly discusses the volatility research of high frequency data based on GARCH model, selects the maximum and minimum of 15-minute income data of CSI 300 index in Chinese stock market as the research object, and uses GARCH model to model and analyze. The first chapter introduces the current situation of volatility research at home and abroad. The second chapter systematically introduces the basic theory of volatility econometric model and the main characteristics of volatility, and carries on the theoretical comparison among various models, in order to reveal the different characteristics of each model. In the third chapter, the GARCH model is used to model the return series, and the volatility aggregation, the spike and the thick tail and the asymmetry of volatility are analyzed. The conclusion shows that the stock market in Shanghai and Shenzhen has obvious volatility, and the data of yield has its own peak and thick tail, and the volatility is concentrated, and the non-normal distribution of the stock market is not normal. Finally, the analysis results of this paper are summarized, and the improvement of the model is put forward.
【学位授予单位】：长春工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：O211.61;F830.9

【参考文献】