Our main goal in this thesis is to analyze the pricing of CDS contracts in a structural-default model with jumps. In contrast to the seminal paper by Black and Cox (1976), where the firm-value process is modelled as a geometric Brownian motion, a closed-form expression of the distribution of the time of default is not known when jumps are incorporated. We consider a model with double-exponentially distributed jump sizes, where the Laplace transform of the time of default can be calculated. The numerical inversion of this transform is said to be ill-posed and consequently we intensively discuss and compare four inversion techniques in a double-precision environment to retrieve default probabilities. It turns out that in particular the Fixed-Talbot Algorithm performs very well in this credit pricing environment. As a result, this Laplace transform Algorithm allows for an extremely fast calibration to market quotes. The model is calibrated to the eighth iTraxx series over a three-month period in the middle of the subprime crisis. Our experiments reveal that the model?s fitting capabilities are excellent with errors far below bid-ask spreads. Finally, we provide a method that stabilizes the parameters over time, based on the relative entropy of two equivalent probability measures by giving up only slight fitting qualities. This leads to a model that is easy to interpret, numerical robust, and capable to explain observed market spread curves.     «

Our main goal in this thesis is to analyze the pricing of CDS contracts in a structural-default model with jumps. In contrast to the seminal paper by Black and Cox (1976), where the firm-value process is modelled as a geometric Brownian motion, a closed-form expression of the distribution of the time of default is not known when jumps are incorporated. We consider a model with double-exponentially distributed jump sizes, where the Laplace transform of the time of default can be calculated. The n...     »