The credit default swap market\u27s determinants, efficiency, and relationship to the stock market

The Credit Default Swap (CDS) market is a rapidly growing market in which participants such as banks and hedge funds actively trade credit risk. The increasing availability of pricing data has made the CDS market a growing area for empirical research. Much of that research focuses on three topics: the determinants of credit default swap spreads, the efficiency of the CDS market, and the relationship among the CDS, stock, and bond markets. This thesis is primarily concerned with examining the efficiency of the CDS market. However, in doing so, I contribute to all three strands of empirical research. By first estimating a regression model based on variable implied by structural models, I find that a rating-based index is the single best predictor of CDS spread changes followed by firm-specific equity market information. Second, I conduct an event study that, to the best of my knowledge, is the first investigation of the CDS market\u27s reaction to earnings announcements. This event study incorporates a comparative framework in which both the CDS and the stock markets\u27 responses to earnings announcements and credit rating changes are considered. I find that both markets have statistically significant reactions to earnings announcements and credit rating changes and both markets anticipate these informational events up to 90 trading days prior to announcement. However, the presence or absence of post-announcement abnormal performance is sensitive to model choice, and thus the null hypothesis that the CDS market is efficient cannot be rejected. Finally, a vector autoregressive framework is implemented to explore the relationship between the CDS market and the stock market. I find that stock returns lead CDS spread changes, and this finding prevails even under conditions of credit deterioration. However, while the VAR results suggest that the stock market is relatively more efficient, the overall explanatory power of the VAR model combined with subsequent variance decompositions, cointegration analysis, and the findings from the event study lead me to the ultimate conclusion that the CDS market is, on the whole, an efficient one