Stress scenario development for commercial bank in Czech Republic

The primary focus of this thesis is to introduce stress testing process. It begins with the past development, the current situation and the legislative part of this field. The main part of the thesis discusses the methodology of the stress scenario design as the scenario itself is the required outcome of the whole process. First, it is necessary to determine the risk factors the company is sensitive to, which are then subjects to forecast in order to obtain the scenarios. Then, different approaches follow to ensure the scenario is severe but plausible because that is what the regulation requires. For this purpose, the severity function approach (SFA) is used along with the probability distribution function obtained from risk factors forecasting. Macroeconomic variables are chosen as risk factors in this thesis; therefore, Bayesian Vector Autoregressive model with proper prior distribution takes place as a forecasting tool. The theoretical background is demonstrated in the second part of the thesis on a real dataset from the Czech Republic from a credit risk management of a commercial bank perspective.