Contagion and interdependence on the CDS market: a Bayesian Factor Model approach

Absztrakt (kivonat)

The Credit Default Swaps played a highlighted role in the last financial and debt crises, therefore several studies have been done to understand their dynamics. The Credit Default Swaps are functioning as quasi-insurances on financial markets, thus the sovereign CDS spreads can be used as indicators of the default risk of the corresponding countries. After the Lehman Brothers has collapsed and the Greek debt crisis has revealed, the literature became focused on the contagion and interdependence phenomena. These concepts show how the financial markets are connected, and how these linkages are changing after different shocks occur. Studying contagion and interdependence is relevant in the sovereign credit markets as they explicitly represent the default risk of the sovereigns, therefore, findings can have implications on monetary policies regarding measures that could mitigate the vulnerability of a sovereign when other countries are experiencing crisis periods. Our contribution to the existing literature is the application of an advanced methodology, namely the estimation of a Bayesian Dynamic Factor model with time-varying parameters and stochastic volatility on a large dataset containing credit spreads from 2005 to 2018 for 39 countries. In line with the existing literature, we find a high degree of commonality between the CDS spreads, while we estimate a significant global and European factor with clear economic interpretation, but no regional factor for the Latin American and Asian countries. We observe strong interdependence on the global and European credit risk market, while contagion is also present for a certain group of countries. We find contagion for Spain, Italy, and Portugal as their connectedness to the global factor is increased permanently after the global financial crisis, while a significant increase in the linkage between Eurozone member states is present during the European sovereign debt crisis.