Hodnocení úvěrového rizika v mezinárodním obchodě - srovnání modelu EGAP, a. s., a komerčních bank

The dissertation compares the export credit rating model of the national Export Guarantee and Insurance Agency EGAP with models applied by selected Czech banks. The first part of the dissertation presents a summary of credit risk theory. It depicts the main principles of lending and its risks. The dissertation further describes the factors that influence credit risk and the methods of its modelling. While mathematical risk models project the expected loss as well as its sensitivity to the risk factors, the focus of this thesis lies in qualitative models which set a normalized scale for probability of default, the so called credit rating models. The main contribution of the dissertation lies in the survey carried out among four Czech banks belonging to owners from various countries, from which we get an overview of their rating models. It follows from the gathered information that their models are based on financial indicators when rating the buyers/exporters. The models are also considerably amended by non-financial factors whose importance in certain cases rose following the recent financial crisis. The agency EGAP insures business activities abroad and therefore its model takes into account also specific factors related to the destination country. The main difference between the models of EGAP and the examined banks lies in the method of creation and validation: EGAP does not dispose of sufficient amount of business case studies, so that it has to rely on external consulting services when setting up and validating the model. The dissertation concludes that while all rating models are composed of similar risk factors highlighting past financial indicators of the financed business, each analysed rating model differs significantly in the specific database of business cases that were used to construct the model, depending on the availability of data to the bank/insurer. The conclusion that can be drawn from this fact is that the main factor for successful prevention of future failures of the credit rating models will be the extent of the credit assessment database which will be used for the construction of the respective rating model.