Abstract
Bank solvency was a major issue during the financial crisis of 2007–2009, but bank credit default swap (CDS) spreads were almost always below nonbank CDS spreads. What is the reason for this gap? Are banks perceived to be less risky? This study empirically decomposes CDS premia for 45 major banks and 167 large industrial firms from Europe and the US. It turns out that expected losses are usually somewhat lower for banks than for nonbanks, but expected losses contribute relatively little to the observed CDS premia. CDS spreads for banks and nonbanks differ mainly because market participants require a lower compensation for bearing bank credit risk. The quite persistent difference in the credit risk premia for banks and nonbanks disappears only temporarily during the crisis.
Acknowledgement
I would like to thank Phillip Hartmann for his support and Helmut Elsinger, Martin Scheicher and Esther Segalla for their useful comments.
Notes
1 In 2009, the ISDA introduced some changes to standardize CDS contracts even further and to enhance liquidity even more in its ‘Big Bang Protocols’. Markit (Citation2009) provides details.
2 Arora et al. (Citation2012) find that counterparty credit risk has only a vanishingly small effect on CDS premia.
3 Asian companies were excluded because the CDS market for Asian firms was relatively small and illiquid compared to markets for European and US firms in the first years of the sample period (Remolona and Shim, Citation2008).
4 The number of firms differs in the subsamples because some firms drop out in certain subsamples, either because of default (i.e. Lehman) or because of missing data. The composition of the full sample and the subsamples is available upon request.
5 See also Berg (Citation2010) for a discussion of this point.