The Complete Picture of Credit Default Swap Spreads - A Quantile Regression Approach

We study the determinants of Credit Default Swap (CDS) spreads through quantile regressions. The results indicate that CDS spreads are strongly determined not only by traditional theoretical variables, such as the implied volatility and put skew, but also by illiquidity costs. However, contrary to stocks or bonds, we show that CDS transaction costs should be measured by absolute, rather than relative, bid-ask spreads. The quantile regressions reveal significant heterogeneity in the response of low-risk versus high-risk firms, with both the coefficients on the explanatory variables and the goodness-of-fit of the model increasing with the quantile of CDS premiums. The results obtained with a classical conditional mean regression, as used in most previous empirical studies, are very far from describing the centre of the CDS distribution, being more representative of the higher quantiles. These results are also helpful to understand the credit spread puzzle