Abstract
Multidimensional dependence in financial markets has motivated the conception of copulas as a tool to analyze nonlinear connections. However, the dynamics generating the dependence structure is still considered unchanged, whether during turmoil or stable periods, and the curse of dimensionality prevents researchers from detecting the regime shifting that may connect financial markets. In this paper, we develop a tractable Markov regime-switching C-vine and D-vine under the symmetrized Joe-Clayton copula, capable of detecting lower and upper tail dependencies separately. Application is conducted on twelve government bonds, the U.S. and eleven European bonds belonging to the Eurozone. Results show that the regime-switching copula models explain the dynamics of data dependence better than the single-regime copula, which indicates the presence of a contagion effect. Furthermore, for Eurozone bond markets, the contagion remains in its high state since the global financial crisis of 2008 and European sovereign debt crisis of 2009, with a transmission path from core to stressed countries.
•We develop a tractable Markov regime-switching C-vine and D-vine copulas.•Application is performed on 12 government bonds: the U.S. and 11 European bonds.•The superiority of the two-regime models over the single-regime strongly suggests the presence of a contagion effect.•The contagion remains in its high state since the global financial and European sovereign debt crises.•The transmission path is from core to stressed countries.