Abstract
This study introduces a novel copula class, referred to as the distorted GAB copula (hereafter, dGAB copula), as an alternative to the Gaussian copula, which has shown limitations in capturing tail dependence. Much like the Gaussian copula, the dGAB copula can be uniquely determined by its bivariate marginal copulas and offers effective tail dependence modeling capabilities. To demonstrate its practical applicability, we showcase its use in the valuation of basket default swaps. Furthermore, we propose a parameter estimation approach based on the EM algorithm tailored to the dGAB copula.
https://www.tandfonline.com/doi/full/10.1080/01605682.2024.2305646
