Tractable multi-firm default models based on discontinuous processes

The aim of this thesis is to thoroughly study structural default models based on jump-diffusion processes. Jump-diffusion models were first proposed by Zhou (2001), who also showed that these models have several desirable properties, most important, positive short-term spreads. On the other side, these models turned out to be complicated to implement, as the distribution of the running minimum of a jump-diffusion process is not known. We show that an efficient implementation of jump-diffusion models is possible and even fast enough to allow a calibration to market data. Moreover, we derive new theoretical results and generalize the model to jointly consider different companies.