Testing and Estimation of Structural Breaks in Fractional Cointegration

分数协整中结构断裂的测试和估计

• 批准号: 452510572
• 负责人: Professor Dr. Philipp Sibbertsen
• 金额: --
• 依托单位: Institut für Statistik
• 依托单位国家: 德国
• 项目类别: Research Grants
• 资助国家: 德国
• 项目状态: 未结题
• 来源: https://gepris.dfg.de/gepris/projekt/452510572?language=en
• 关键词: Testing; Estimation; Structural; Breaks; Fractional

The aim of the project is to develop and empirically apply tests for time-varying partial fractional cointegration. If one interprets a fractional cointegration relationship as a long-term stable equilibrium, it can be assumed that this also changes with changing economic and political conditions. The change can be that the cointegration vector changes over time, but the actual equilibrium relationship is stable, or that fractional cointegration is only present at specific times. Both have not been considered by the literature so far, so that the aim of this project is to develop methods that test for such temporal changes in the cointegration relationship and then allow them to be modeled and interpreted.On the one hand, the test for fractional cointegration by Hassler and Breitung (2006) is combined with approaches by Davidson and Monticini (2010) to allow for breaks in the cointegration relationship. On the other hand, a likelihood ratio-based approach is used which tests for multiple structural breaks in the cointegration vector in a cointegrated multivariate time series system.Furthermore, the aim of this project is to see whether the occurrence of fractional cointegration relationships is driven by a background variable. In order to do this the cointegration residuals are modeled as a Markov switching process.With the help of these theoretical methods, the relationship between fractional cointegration and market integration in the Eurozone is examined. It is considered which economic knowledge can be gained by applying these methods to the correlation between equity and bond yields (so-called “flight-to-quality” effects), as well as to dependency structures between the volatilities of different asset classes, and to international financial market volatilities .

该项目的目的是开发和经验应用测试随时间变化的部分分数协整。如果将分数协整关系解释为长期稳定均衡，则可以假设这也会随着经济和政治条件的变化而变化。这种变化可以是协整向量随时间变化，但实际的均衡关系是稳定的，或者是分数协整只存在于特定的时间。到目前为止，这两种情况都没有被文献考虑过，因此本项目的目的是开发测试协整关系中这种时间变化的方法，然后允许它们被建模和解释。Hassler和Breitung（2006）的分数协整检验与Davidson和Monticini（2010）的方法相结合允许协整关系的中断。另一方面，本研究利用似然比方法来检验多元时间序列协整系统中协整向量的多重结构突变，并进一步探讨分数阶协整关系的发生是否受背景变量的驱动。为了做到这一点，协整残差被建模为一个马尔可夫转换过程，借助这些理论方法，分数协整和市场一体化之间的关系，在欧元区进行了检查。人们认为，通过将这些方法应用于股票和债券收益率之间的相关性（所谓的“向质量转移”效应），以及不同资产类别的波动率之间的依赖结构和国际金融市场波动率，可以获得哪些经济知识。