房地产市场非对称及厚尾相依性研究

发布时间：2018-03-06 15:45

本文选题：房地产市场　切入点：Copula-GARCH模型　出处：《华中科技大学》2016年博士论文　论文类型：学位论文

【摘要】：2015年,全球经济维持“弱增长”格局,仍延续曲折性与脆弱性并举的调整恢复期。货币基金组织最新发布的《世界经济展望》指出,全球经济当前正面临着不可忽视的下行风险,这会直接影响全球经济未来的发展。随着世界经济一体化的进程,资本的自由流动和国际间贸易的联系益发密切,全球金融市场及经济体之间已呈现出愈发明显的相依性,区域内金融市场的波动也日趋受到区域外其他金融市场不确定性的影响。金融市场间的时变相依性的存在不可避免地也会对投资组合的风险管理与资产定价产生影响,并对政策制定者提出更高的挑战。能否准确的刻画和预测全球金融市场间的动态相依性特征,对引导资金的跨市场流动和资源配置、市场参与主体的投资策略制定、监管机制的政策制定及实施等问题均具有重要的现实意义。同国际股票市场的相依性一样,国际房地产市场间相依性结构也有动态性、非对称性以及尾部相依的特征。传统的金融工具之间的整体相依性通常是通过皮尔森线性相关性系数来刻画。但是,近年来,这个方法受到来自学术界和业界研究人员大量的批评,根本原因在于它无法刻画非线性的相依性结构。而Copula模型能够很好的弥补线性相关系数的缺陷,也能够同时对动态性、非对称性和尾部相依结构进行建模。因此,本文在Copula框架下,融合GARCH模型,对全球多个房地产市场的动态、非对称相依性和尾部相依性展开研究。首先,本文在国际上研究相依性的先进技术框架下,通过结合Copula模型和ARMA-GJR-GARCH模型,研究了包括房地产非对称相依性在内的三种非对称性。本文分别采用ARMA-GJR-GARCH模型来捕捉房地产市场的非对称波动,采用偏斜t分布拟合非对称的边际分布,最后通过GARCH模型和动态Copula模型结合,来刻画房地产市场非对称的时变相依性结构。实证结果表明,美国和英国房地产市场之间的相关性最强,并且呈显著增强趋势。其次,针对房地产市场的非对称相依性特征,本文基于三种动态Copula模型,结合风险管理模型,在动态Copula-VaR框架下预测房地产投资组合的在险价值,揭示非对称相依性结构的风险管理意义。实证结果表明,如果投资者忽略极值联动会低估投资组合的预期风险,导致更大的损失；同时对称的t模型得到的房地产资产投资组合的VaR值也比非对称的旋转Gumbel模型得到的VaR要更低。因此,如果在椭圆相关性的假设下会导致风险管理者对风险的低估。第三,基于极值理论和尾部相依性,分析极值事件(金融危机)对房地产市场间相依的影响。通过分析次贷危机前、危机期和危机后七个国际房地产市场的尾部相依性,考察了金融危机对房地产市场相依性、及房地产与一般金融市场间相依性的影响。实证结果显示,2008年后,几乎所有的国际房地产市场之间都从尾部不相依转变成了尾部相依。而这种较强的相依性并没有随着全球经济的逐渐恢复而变弱,由此揭示金融危机对全球房地产证券市场的影响是非常深远的。最后,本文基于动态SJC Copula-GARCH-t模型,对中国房地产和股票市场间的动态尾部相依性进行了建模和分析。实证结果显示,中国房地产-股票市场间的尾部相依性一直保持在高位水平,并且受到金融危机和欧债危机的影响较为有限。这与其他国家的房地产-股票市场间的实证结果有非常大的差异。可以说,来自国际市场的极值事件对中国市场的冲击并不强烈。
[Abstract]:In 2015, the global economy remained weak growth pattern, continue twists and vulnerability and the adjustment of the recovery period. The IMF's latest world economic outlook < > pointed out that the global economy is facing downside risks can not be ignored, which directly affects the development of the global economy with the world economy in the future. The process of integration, the free flow of international capital and trade links between the global financial market more closely, and the economy has shown more obvious dependence, regional financial market volatility is also affected by other regional financial market uncertainties. The time-varying dependence between financial markets are inevitably the also on the portfolio risk management and asset pricing impact, and put forward higher challenge to policy making. The ability to accurately describe and predict the global financial markets The characteristics of the dynamic dependencies, to guide the fund flow across the market and the allocation of resources, develop market participants investment strategy, the regulatory mechanism of the policy formulation and implementation and other issues have important practical significance. As the dependence of the international stock market, international real estate market dependence between structure and dynamic characteristics. Asymmetric and tail dependence. The overall dependence between traditional financial instruments of is usually described by Pearson linear correlation coefficient. However, in recent years, this method suffers from academic and industry researchers a lot of criticism, the fundamental reason is that the dependence of the structure. And it can not describe the nonlinear Copula model can good to make up for the defect of linear correlation coefficient, is also on the dynamic, asymmetric and tail dependence structure is modeled. Therefore, in this paper, under the framework of Copula, fusion The GARCH model, the dynamic of the global number of the real estate market, asymmetric dependence and tail dependence is studied. Firstly, this paper studies the advanced technology framework of dependence in the world, by combining the Copula model and ARMA-GJR-GARCH model are studied, including the real estate of the asymmetric three asymmetric dependence,. Asymmetric Volatility we use the ARMA-GJR-GARCH model to capture the real estate market, the marginal distribution of the skew t distribution fitting asymmetric, finally through the combination of GARCH model and dynamic Copula model to describe the real estate market asymmetric time-varying dependence structure. The empirical results show that the correlation between the United States and the British real estate market, and showed a significant increasing trend. Secondly, the asymmetric dependence on the real estate market characteristics, the three kinds of dynamic Copula model based on the combination of risk management in the model. Under the framework of Copula-VaR dynamic prediction of real estate portfolio value at risk, reveal the asymmetric dependence structure of the meaning of risk management. The empirical results show that, if investors ignore the expected risk of extreme linkage portfolio could be underestimated, leading to greater losses; at the same time, the symmetrical t model get a real estate asset portfolio value VaR non symmetric rotating Gumbel models get VaR to lower. Therefore, if the relationship between elliptic under the assumption that will lead to the risk management of underpricing of risk. Third, the dependence of extreme value theory and based on the analysis of extreme value event tail, (financial crisis) on the real estate market dependence effect. Through the analysis of the subprime crisis. The period of crisis and crisis after the end of seven the international real estate market dependence, effects of financial crisis on the real estate market dependence, and real estate and financial markets according to the phase of Impact. The empirical results show that, after 2008, almost all of the international real estate market is not dependent from the tail into the tail dependence. This dependence is not strong with the gradual recovery of the global economy becomes weak, thus revealing the impact of financial crisis on the global real estate securities market is very far-reaching. Finally, the dynamic SJC Copula-GARCH-t model based on dynamic tail China real estate and the stock market's dependence is modeled and analyzed. The empirical results show that the tail of China real estate stock market dependence has been maintained at a high level, and are influenced by the financial crisis and the European debt crisis is relatively limited empirical. This and other countries of the real estate and stock market results have very big difference. It can be said that no strong impact of extreme events from the international market for China market.

【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F299.23;F832.51;F224

【相似文献】