基于SJC Copula模型的银行业与房地产业动态相依性及其结构突变

Dynamic Dependency Structure and Its Structure Break Point between China's Real Estate Industry and Bank Industry Based on SJC Copula Model

引入SJC Copula模型来刻画中国房地产业与银行业在2000~2013年的的动态相依性,并进一步分析这种动态相依性的结构突变特征。研究结果表明:SJC Copula函数相比常见的其他Copula函数能更好地刻画房地产和银行业之间的动态相依结构;房地产业与银行业之间存在具有显著持续性的上(下)尾相依性,而且随着收益率的变化而呈现出非对称和非线性变化特征,主要表现为随着收益率的上升,条件上尾(下尾)相关系数将减小,但上尾相依性系数的下降幅度比下尾相依性系数要大,这说明房地产业和银行业之间的相依性的变化显著依赖于过去的收益的变化并且存在反向关系。房地产业与银行业之间的上(下)尾时变相关系数均具有"自相关、有偏、ARCH效应"的非正态分布特征,而且存在多个结构突变点;这些突变点发生日期附近往往伴绥着我国重大房地产金融调控政策的出台,具有明显的政策效应,但不具有持久性。

This paper builds SJC copula model to characterize the dynamic dependency structure and its structural break point between real estate industry and bank industry based on their daily index yield data from 2000 to 2013.The results show that SJC copula can better depict the dependency between the two industries compared with other copula.There are significantly persistent upper(lower)tail dependence between Real estate and banking sector.Those upper(lower)tail dependence shows asymmetrical and non-linear variation with the change in yield.With the rise in yields,conditions upper(lower)tail correlation coefficient will decrease,but upper tail dependence decline larger than lower tail lower tail dependence.This shows that the dependency change depends significantly on the return of the past and there is a reverse relationship.The higher the volatility of the return of the past,the closer the relationship between the two industries is,and vice versa.These upper(lower)tail dependence shows some non-normal distribution characteristics such as Autocorrelation,Biased and ARCH effect.At the same time,there are some structural break points among these time-varying linkages between the real estate and bank sector.The date of the occurrence of these structural mutations is often accompanied by major real estate financial control policies which has obvious policy effects.These policies do not have persistent effects.