变结构特征对GARCH模型波动持续性的影响  

Study on Effect of Variable Structure Characteristics on the Volatility Persistence of GARCH Model

作　　者：孙晶[1] 江孝感[2] 

机构地区：[1]南京工业大学浦江学院商学院,江苏南京210034 [2]东南大学经济管理学院,江苏南京211189

出　　处：《江苏科技信息》2016年第18期62-66,共5页Jiangsu Science and Technology Information

摘　　要：文章认为忽略二维金融时间序列中存在的变结构现象来构建GARCH模型,会导致模型波动持续性的高估。在此基础上根据以往的经验分析模型的协同持续关系,构建协同持续向量,在金融时间序列各分量波动平稳而组合的向量金融时间序列波动持续的情况下会表面上消除了方差的波动持续性,但本质上不但没有降低,反而提高了组合的波动持续性。在资产投资组合分析中,如果忽视各资产时间序列中的变结构特征而构建资产投资组合,不但无法实现风险的规避,反而会提高组合的金融风险。笔者认为,进行变结构特征的存在性分析是对多维金融时间序列构建GARCH模型以真实描述其波动持续性的前提,也是资产投资组合分析构建长期资产投资组合以实现风险规避的前提。This paper argues that to construct GARCH model ignoring the existing variable structure in two-dimensional financial time series would lead to an overestimation of continuing volatility of model. On this basis, cooperative continuing relationship of model is analyzed based on past experience and cooperative continuing vector quantity. When each component of financial time series fluctuates smoothly and steadily and the vector of combined financial time series fluctuate continuously, the persistence in volatility of variance will be eliminated in surface. It is not reduced substantially; combined persistence in volatility is increased. In the analyses of asset portfolio, it will not achieve risk aversion to build asset investment portfolio neglecting variable structural features of each asset time series, while it will increase the combined financial risk. This paper argues that analysis of the existence of variable structure characteristics is the premise to build GARCH model using multidimensional financial time series and describe the volatility persistence in reality. And it is the premise to analyze asset investment portfolio, to build long-term asset investment portfolio and to achieve risk aversion.

关 键 词：GARCH 变结构 波动持续性 协同持续性 

分 类 号：O211.6[理学—概率论与数理统计]