The implementation of the Basel II capital adequacy framework promoted internally modelled risk parameters and allowed banks to build their own models. The recent crisis pointed at the gaps in the Basel II Accord, seeing banks having trouble to deal with lack of liquidity and higher default rates. The minimum regulatory capital held by the banks turned out to be insufficient and banks started looking for other techniques to better quantify the risks they are exposed to. Model accuracy is a key objective to meet the capital adequacy requirements while facing severe economic conditions. The purpose of this thesis is to suggest a new approach to credit modelling. Data envelopment analysis (DEA) can overcome some the difficulties that the banks deal with. The key opportunity in using DEA and its modifications is in the fact that this method does not require prior information about the classification between good and bad units and only requires financial and other data about the client in question. This thesis analyses the performance of DEA applied on a real world portfolio of corporate loans compared to the two standard methods used in the banking sector. Logistic regression is the most popular method, having few restrictions and providing output in the form of a probability of default. The second method is the discriminant analysis giving similar results to the logistic regression but being based on more assumptions. The model is validated by comparing the model output with the actual status and its predictive power evaluated.
| Identifer | oai:union.ndltd.org:nusl.cz/oai:invenio.nusl.cz:191818 |
| Date | January 2014 |
| Creators | Fialová, Zuzana |
| Contributors | Jablonský, Josef, Lukáš, Ladislav, Brezina, Ivan |
| Publisher | Vysoká škola ekonomická v Praze |
| Source Sets | Czech ETDs |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/doctoralThesis |
| Rights | info:eu-repo/semantics/restrictedAccess |
Page generated in 0.0013 seconds