中国股市量价关系的长记忆性及非线性时变相关研究
发布时间：2019-05-19 09:44
【摘要】：资本市场中交易量与价格之间的关系奠定了技术分析在现代证券投资的核心地位,而混合分布假设则为量价关系的研究提供了理论依据.相较于实务分析,量价间理论分析成果尚有可改善之处. 本文基于长记忆的分形思想、多元金融时序GARCH波动建模以及度量非线性时变相关的Copula理论,研究了量价间的深层关系,并辅之以实证与分析,做了如下两方面的主要工作： 一方面,现有研究主要是把交易量作为外生变量加入到收益率波动方程中研究量价间线性关系,本文则同时以交易量及收益率为研究主体,并考虑到波动的持续性.运用了二元GARCH及Copula理论的研究方法,深入、系统研究了量价间的相关关系. 运用长记忆金融时序检验方法,针对上证指数对数收益及交易量序列进行了长记忆的检验.结果支持两水平序列不存在长记忆性,但波动序列表现出较强持续性.基于长记忆的波动模型分别对交易量及收益率序列建立了长记忆的波动模型,并与短记忆模型进行了比较.模型结果表明量价时序波动均表现出长记忆性,这表明短记忆模型对量价时序波动描述不足.同时量价两序列具有相近的分形参数,这表明两序列具有相同的持续性.利用多元GARCH模型思想,运用二元GARCH模型分析了中国股市交易量与收益率间的关系,模型结果表明BEKK模型最优,量价间存在显著正相关. 另一方面,引入能度量非线性、尾部相关、时变相关的Copula理论分析中国股市量价关系,提出了能度量量价两时序异方差、长记忆、时变、非线性尾部相关的SJCTVPCopula-FIGARCH-t模型,通过对上证指数对数收益率及交易量两序列的实证分析表明该模型优于其他现有模型.结果表明上证指数在样本期内交易量与对数收益率本身更贴近t分布而非正态分布,均存在长记忆性,且长记忆性参数近似相等.量价间存在非线性上尾相关、相关性表现出较明显时变的正相关.
[Abstract]:The relationship between trading volume and price in capital market lays the core position of technical analysis in modern securities investment, and the mixed distribution hypothesis provides a theoretical basis for the study of quantity-price relationship. Compared with the practical analysis, the results of the theoretical analysis between quantity and price can still be improved. Based on the fractal idea of long memory, the modeling of multiple financial time series GARCH volatility and the Copula theory of measuring nonlinear time-varying correlation, this paper studies the deep relationship between quantity and price, supplemented by empirical and analysis. The following two main work has been done: on the one hand, the existing research mainly adds the trading volume as an exogenous variable to the volatility equation of return to study the linear relationship between quantity and price, while this paper takes the trading volume and rate of return as the main body of the study. And take into account the persistence of fluctuations. By using the research methods of binary GARCH and Copula theory, the correlation between quantity and price is studied systematically. The long memory financial timing test method is used to test the logarithmic return and trading volume sequence of Shanghai Stock Exchange Index. Results there was no long memory in the two level sequences, but the fluctuation sequences showed strong persistence. Based on the long memory fluctuation model, the long memory fluctuation model is established for the trading volume and the rate of return sequence, and compared with the short memory model. The results of the model show that the temporal fluctuation of quantity and price shows long memory, which indicates that the short memory model is not enough to describe the fluctuation of quantity and price. At the same time, the two sequences have similar fractal parameters, which indicates that the two sequences have the same persistence. Based on the idea of multivariate GARCH model and binary GARCH model, the relationship between trading volume and return in Chinese stock market is analyzed. The results of the model show that the BEKK model is optimal and there is a significant positive correlation between quantity and price. On the other hand, Copula theory, which can measure nonlinear, tail correlation and time-varying correlation, is introduced to analyze the quantity-price relationship in Chinese stock market, and a SJCTVPCopula-FIGARCH-t model which can measure 02:00 order heteroscedasticity, long memory, time-varying and nonlinear tail correlation is proposed. Through the empirical analysis of the logarithmic rate of return and trading volume of Shanghai Stock Exchange Index, it is shown that the model is superior to other existing models. The results show that the trading volume and logarithmic rate of return of Shanghai Stock Exchange Index are closer to t distribution than normal distribution during the sample period, and all of them have long memory, and the long memory parameters are approximately equal. There is a nonlinear upper-tail correlation between quantity and price, and the correlation shows obvious time-varying positive correlation.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51
本文编号：2480617
[Abstract]:The relationship between trading volume and price in capital market lays the core position of technical analysis in modern securities investment, and the mixed distribution hypothesis provides a theoretical basis for the study of quantity-price relationship. Compared with the practical analysis, the results of the theoretical analysis between quantity and price can still be improved. Based on the fractal idea of long memory, the modeling of multiple financial time series GARCH volatility and the Copula theory of measuring nonlinear time-varying correlation, this paper studies the deep relationship between quantity and price, supplemented by empirical and analysis. The following two main work has been done: on the one hand, the existing research mainly adds the trading volume as an exogenous variable to the volatility equation of return to study the linear relationship between quantity and price, while this paper takes the trading volume and rate of return as the main body of the study. And take into account the persistence of fluctuations. By using the research methods of binary GARCH and Copula theory, the correlation between quantity and price is studied systematically. The long memory financial timing test method is used to test the logarithmic return and trading volume sequence of Shanghai Stock Exchange Index. Results there was no long memory in the two level sequences, but the fluctuation sequences showed strong persistence. Based on the long memory fluctuation model, the long memory fluctuation model is established for the trading volume and the rate of return sequence, and compared with the short memory model. The results of the model show that the temporal fluctuation of quantity and price shows long memory, which indicates that the short memory model is not enough to describe the fluctuation of quantity and price. At the same time, the two sequences have similar fractal parameters, which indicates that the two sequences have the same persistence. Based on the idea of multivariate GARCH model and binary GARCH model, the relationship between trading volume and return in Chinese stock market is analyzed. The results of the model show that the BEKK model is optimal and there is a significant positive correlation between quantity and price. On the other hand, Copula theory, which can measure nonlinear, tail correlation and time-varying correlation, is introduced to analyze the quantity-price relationship in Chinese stock market, and a SJCTVPCopula-FIGARCH-t model which can measure 02:00 order heteroscedasticity, long memory, time-varying and nonlinear tail correlation is proposed. Through the empirical analysis of the logarithmic rate of return and trading volume of Shanghai Stock Exchange Index, it is shown that the model is superior to other existing models. The results show that the trading volume and logarithmic rate of return of Shanghai Stock Exchange Index are closer to t distribution than normal distribution during the sample period, and all of them have long memory, and the long memory parameters are approximately equal. There is a nonlinear upper-tail correlation between quantity and price, and the correlation shows obvious time-varying positive correlation.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51
【参考文献】
相关期刊论文 前10条
1 李付军,达庆利;中国股市量价波动性关系的实证分析[J];东南大学学报(自然科学版);2005年02期
2 王春峰,张庆翠,李刚;中国股票市场收益的长期记忆性研究[J];系统工程;2003年01期
3 韦艳华,张世英;金融市场的相关性分析——Copula-GARCH模型及其应用[J];系统工程;2004年04期
4 易文德;;基于Copula函数模型的股市交易量与股价相依关系[J];系统工程;2010年10期
5 韦艳华;齐树天;;亚洲新兴市场金融危机传染问题研究——基于Copula理论的检验方法[J];国际金融研究;2008年09期
6 李文君;尹康;;多元GARCH模型研究述评[J];数量经济技术经济研究;2009年10期
7 杨渺,杨代若;A、B股指数波动的Granger因果关系分析[J];数理统计与管理;2003年01期
8 韦艳华;张世英;;多元Copula-GARCH模型及其在金融风险分析上的应用[J];数理统计与管理;2007年03期
9 夏天;;基于CARR模型的交易量与股价波动性动态关系的研究[J];数理统计与管理;2007年05期
10 李双成;王红霞;;中国股票市场交易量与价格波动关系实证研究[J];数学的实践与认识;2008年12期
，本文编号：2480617
本文链接：https://www.wllwen.com/guanlilunwen/zhqtouz/2480617.html
最近更新
教材专著
