6 Conclusion
In this paper, we show that liquidity tail risk in the CDS market significantly comoves with CDS spreads. We make use of a dynamic copula model to estimate the upper tail dependence between a CDS contract’s idiosyncratic bid-ask spreads and illiquidity in the CDS market (i.e., liquidity tail betas). In panel regressions, we then regress the CDS spreads of our sample firms on the contracts’ liquidity tail betas and various controls for the firms’ default risk. The results that we find have important implications for risk managers and investors that enter the CDS market as net protection sellers like insurers and pension funds. Our results provide ample evidence for the presence of time-varying liquidity tail risk in the CDS market. Liquidity tail risk spiked across our full sample during the financial crisis with peaks in liquidity tail risk appearing shortly after the bailout of AIG and at the start of 2009. Moreover, monthly CDS spreads comove significantly with liquidity tail betas with protection sellers demanding a premium for bearing liquidity tail risk.