Score-driven Systemic Risk Signaling for European Sovereign Bond Yields and CDS Spreads

In this chapter, we investigate the robustness of systemic risk indicators based on alternative but related data sources. Our focus is on indicators based on the generalized autoregressive score methodology of Creal et al. In particular, we provide new empirical results for joint default probabilities based on score-driven models using historical government bond yield data versus CDS spread data. Although CDS data are the more common choice in the empirical literature for measuring sovereign systemic risk, the typical time span for obtaining reliable data is rather short. Bond yield data, by contrast, allow us to go further back in time and therefore to test the adequacy of systemic risk measures over a much longer historical period.