MODELING OPERATIONAL RISK LOSS DATA USING EXTREME VALUE THEORY
| dc.contributor.author | Pillay, Premlin | |
| dc.date.accessioned | 2011-10-24T13:32:03Z | |
| dc.date.available | 2011-10-24T13:32:03Z | |
| dc.date.issued | 2011-10-24 | |
| dc.description | MBA thesis - WBS | en_US |
| dc.description.abstract | The quantification of operational risk has recently become an explicit requirement for the calculation of capital adequacy for financial institutions. The modeling of operational risk data is difficult due to the infrequency and severity of the loss events. Practitioners of operational risk have generally used the traditional approach in using the lognormal distribution to model this type of heavy tailed data. This project report attempted to determine whether using extreme value theory to model operational risk data is a better approach than using the lognormal approach. The project consists of the fitting of the extreme value theory distribution and the lognormal distribution to a loss dataset and the comparison thereof to determine the better approach. The main finding of the research was that using extreme value theory a better fit to the dataset was obtained than by using the lognormal approach | en_US |
| dc.identifier.uri | http://hdl.handle.net/10539/10648 | |
| dc.language.iso | en | en_US |
| dc.subject | Operational risk | en_US |
| dc.title | MODELING OPERATIONAL RISK LOSS DATA USING EXTREME VALUE THEORY | en_US |
| dc.type | Thesis | en_US |