AbstractsMathematics

Credit risk pricing models assume recovery to be at its \textit{historical} average (historical recovery assumption). However, the effect of this assumption is not completely understood. The heard of this thesis lies in constructing a new pricing model for Credit Default Swaps (CDS), in particularly allowing for negative correlation between recovery and default. This model is denoted as partial differential equations for the CDS legs. By means of an additional Monte Carlo approach we are able to extract continuous implied recovery and default intensity term structures. These structures can then be used to assess the historical recovery assumption. It is in particularly shown that a constant recovery model overestimates the Credit Value Adjustment (CVA) when allowing for perfect negative correlation. While on the other hand it underestimates CVA when it is adjusted for its implied historical average.