6. Sovereign and Bank CDS Spreads During the European Debt Crisis: Laying the Foundation for SMEs’ Financial Distress

Note that the literature on CDSs and its determinants has recently seen growing attention by researchers given the pivotal role played by bank CDSs during the 2007 global financial crisis (see Angeloni and Wolff 2012; Ejsing and Lemke 2011; Demirguc-Kunt and Huizinga 2011; Acharya et al. 2014; Alter and Beyer 2014; Bosma et al. 2012; Gross and Kok 2013). Of particular interest in the field is the contribution by Annaert et al. (2013). The authors explain changes in banks’ CDS spreads through a set of explanatory credit risk variables. The study, inspired by Merton’s theoretical credit risk model (Merton 1974), analyzes 31 European banks. The model includes variables linked to CDS liquidity and variables related to the market and the business cycle. The results show how, during the financial crisis, an increase in credit risk significantly raises the CDS spread. An exhaustive and more complete survey of the literature on CDS is in Augustin et al. (2014).

Indeed, the propagation mechanism of shocks that may affect the decline in bank lending can have several explanations in the literature (Popov and Van-Horen 2015). Some argue that agency costs (Ueda 2012; Dedola et al. 2013) can be viewed as an explanation for the propagation of shocks that reduce the net worth of financial intermediaries; others include capital requirements (Kollmann et al. 2011; Mendoza and Quadrini 2010) and the monopolistic competition that generates countercyclical price-cost margins (Olivero 2010). Alternatively, information asymmetries between banks and investors can be exacerbated by adverse shocks to banks’ net worth, reducing lending to the private real sector.

For the sake of clarity, the stationarity of a series might depend on the choice of the length of the period of observation. The majority of papers that found evidence of non-stationarity usually included the period of high uncertainty and stress as the years 2006–2007, which were characterized by a notable increase in CDS premiums and an important destabilization of the financial system. Additionally, the literature often provides support for the hypothesis of non-stationarity of banks’ CDSs. Therefore, it is pivotal to test for the possible existence of unit roots in our variables. In fact, if banks’ CDSs and sovereign CDSs were integrated in the same order, the analysis of causality could be based on a study of the co-integration of the two series.

Identified as in Avino and Cotter (2014).

Regarding banks’ CDSs, indeed, the hypothesis of stationarity is weakly rejected at a 5.7% probability in the restricted version of the sample.

Indeed, it is not surprising that daily observations can be correlated with their first lag.