This paper proposes a novel asymmetric copula based upon bivariate split normal distribution. This copula can change correlation signs of its upper and lower tails of distribution independently. As an application, it is shown by the rolling maximum likelihood estimation that the EU periphery countries changed sign of the lower tail correlation coefficient from negative to positive after the sovereign debt crisis started. In contrast, Germany had negative stock-bond correlation before and after the crisis.
翻译:本文基于双变法分法正常分配,提出了一个新的非对称交织物。 这种交织物可以独立改变其上尾和下尾分配的关联性标志。 作为应用程序,通过滚动最大可能性估计,欧盟外围国家可以在主权债务危机开始后将低尾相关系数从负值改为正值。 相反,德国在危机前后都存在负股债券相关关系。