In the current risklab Economic Scenario Generator (risklab ESG) a multi-factor Hull-White model, which is based on the cascade structure proposed in the Zagst model, is implemented to generate interest-rate scenarios. On the one hand, this structure has several advantages, e.g. analytical tractability due to the underlying normal distribution. On the other hand, it also bears a major shortcoming because scenarios with negative interest rates can be observed. In this thesis we give an overview on different alternative short-rate models and analyze their advantages and disadvantages. We then extend one of these models - the Cox-Ingersoll-Ross model (CIR) - to be consistent with the cascade structure of the Zagst model. Finally, we calibrate both, the current and the new model, to market data and perform a comparative analysis under different market conditions. We are going to see that when interest rates are at a normal level, comparable term structures can be detected, while low interest rates lead to a better performance of the CIR model.      «

In the current risklab Economic Scenario Generator (risklab ESG) a multi-factor Hull-White model, which is based on the cascade structure proposed in the Zagst model, is implemented to generate interest-rate scenarios. On the one hand, this structure has several advantages, e.g. analytical tractability due to the underlying normal distribution. On the other hand, it also bears a major shortcoming because scenarios with negative interest rates can be observed. In this thesis we give an overview...     »