Disentangling the Information Content of Government Bonds and Credit Default Swaps: An Empirical Analysis on Sovereigns and Banks

We propose a multi-factor Gaussian model to analyze the dynamics of sovereign bond yields, as well as sovereign and banks CDS quotes. This paper has three objectives (all of them with relevant implications from a supervisory perspective): (1) disentangling the credit risk component of sovereign bonds from the interest rate component; (2) exploring the sovereign CDS-bond basis, i.e., the difference between sovereign CDS quotes and the corresponding bond yields; (3) inferring fromCDS quotes the idiosyncratic component of a bank credit risk and analyzing its relation with sovereign risk. We cast the model in a state-space form with linear measurement function. To calibrate the model we consider a maximum likelihood estimation together with a Kalman filter method in which both the gradient vector and the Hessian matrix to be used in the optimization can be computed in closed form.