GARCH模型的改进及其在股市收益波动分析中的应用

发布时间：2018-04-16 17:44

本文选题：波动性 + GARCH模型　；参考：《长春工业大学》2012年硕士论文

【摘要】：长期以来,股票市场收益率波动性的研究历来是金融时间序列研究的关键问题,同时也是每个国家监管机构最关注的权衡目标,因为股票波动性是反映股票价格变化最便捷和最管用的指标之一,同时,与企业投资,财务策划以及消费者的消费行为最为密切相关的。近些年来,我国的股票市场发展迅猛,对股票波动率的变化的度量方法需求也日益强烈。 GARCH模型是为金融数据而专门量体定做的条件异方差模型,其非常适合于股市收益波动性的分析。GARCH模型(广义自回归条件异方差模型)是ARCH模型(自回归条件异方差模型)的推广形式,其充分地说明了资产收益率波动过程。金融时间序列的特点为波动集聚性、非对称性以及尖峰厚尾性。一般地波动集聚性由ARCH模型及其一般形式GARCH模型来刻画,金融时间序列另一个典型的特征是非对称性,GARCH模型中一般都假定残差项是服从正态分布的,不能够表现出这一特点。首先,文章建立了金融时间序列对于波动性非对称性的TAR—GARCH模型(阈自回归GARCH模型),检验股市收益波动对于其自相关性的差异性,以及对于正负信息的差异性。其次,本文引入了虚变量,采用股市收益的虚变量GARCH模型来刻画金融时间序列的非对称性,通过对GARCH模型的改进,更好的反应了金融时间序列的特点及股票波动性的变化,得到了更好的拟合效果。本文将虚变量GARCH模型应用于上证综合指数的波动性研究,拓展了GARCH模型在股票市场上的应用。经研究表明,GARCH模型的改进具有非常重要的理论意义和现实意义,不但可以帮助投资者针对具体情况作出具体分析,而且对于政策的制定者也具有很大的参考价值。最后,又将虚变量GARCH模型和GARCH模型的变体(即非对称GARCH模型)在实例分析中进行对比,更加说明本论文创新点的适用性,即非常适合股票波动性的分析,可以起到非常重要的作用,其意义和对数值本身的研究相比更加显著。
[Abstract]:For a long time, the study of stock market yield volatility has always been a key issue in the study of financial time series, and it is also the most concerned tradeoff goal of every country's regulators.Because stock volatility is one of the most convenient and useful indicators to reflect changes in stock prices, and is most closely related to corporate investment, financial planning and consumer consumer behavior.In recent years, the stock market of our country has developed rapidly, and the demand for measuring the change of stock volatility has become more and more intense.The GARCH model is a conditional heteroscedasticity model tailored for financial data.GARCH model (generalized autoregressive conditional heteroscedasticity model) is a generalization of ARCH model (autoregressive conditional heteroscedasticity model), which fully illustrates the volatility process of asset return.The financial time series is characterized by volatility agglomeration, asymmetry and peak and thick tail.General volatility agglomeration is characterized by ARCH model and its general form GARCH model. Another typical feature of financial time series is asymmetric GARCH model, which is generally assumed that the residual term is obedient to normal distribution, which cannot be shown.Firstly, the paper establishes the TAR-GARCH model of financial time series for asymmetric volatility (threshold autoregressive GARCH model) to test the difference between stock market return volatility and its autocorrelation, as well as the difference of positive and negative information.Secondly, this paper introduces virtual variables, adopts the GARCH model of stock market returns to describe the asymmetry of financial time series. By improving the GARCH model, it better reflects the characteristics of financial time series and the change of stock volatility.A better fitting effect is obtained.In this paper, the virtual variable GARCH model is applied to the volatility study of Shanghai Composite Index, which extends the application of GARCH model in the stock market.The research shows that the improvement of GARCH model is of great theoretical and practical significance. It can not only help investors to make specific analysis according to the specific situation, but also have great reference value for policy makers.Finally, by comparing the variation of virtual variable GARCH model with GARCH model (i.e. asymmetric GARCH model) in the case analysis, this paper demonstrates the applicability of the innovation point of this paper, that is, it is very suitable for the analysis of stock volatility.It can play a very important role, and its significance is more significant than the study of the numerical value itself.
【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】：F224;F832.51

【参考文献】