Sensitive to the need for suitable tools to develop operational risk, in 2001, the BaselCommittee started a series of surveys and statistics regarding operational risks that mostbanks encounter. The idea was to develop and correct measurements and calculationmethods. Additionally, the European Commission also started preparing for the SolvencyII Directive, taking into consideration the operational risk for insurance and reinsurancecompanies.As so, and since Basel and Solvency accords set forth many calculation criteria,our interest in this work, which was developed in parallel with my work on the PhDthesis at bank Audi Lebanon and at the research laboratory of ISFA, Claude BernardUniversity-Lyon, is to elaborate the different quantitative measurement techniques foroperational risk in financial institutions, (particularly in Banks and Insurance companies).We are going to present the associated mathematical and actuarial concepts aswell as a numerical application regarding the Advanced Measurement Approaches, andfocus on the qualitative part of operational risk, mostly the potential for large, unexpectedlosses, either on a per event basis or within a set time period. Hence, pointingout the importance of both qualitative and quantitative measurements and highlightingthe necessity of an operational risk framework. Furthermore, we direct our study workto a more specific type of operational risk which is the estimation risk. We explored suchrisk to some extent by the use of scenario analysis based on expert opinion in conjunctionwith internal data both used to evaluate our exposure to events. In addition, the studywork includes some reflections regarding the measurement of the error induced on theSCR through the estimation error of the parameters. Hence, revealing the importance ofcalling attention to the reflections on assumptions of the calculations.In practice, this work attempts to present the different modeling tools for assessingoperational risk and more particularly, pointing up the consequences of estimation riskbehind. This draws the attention to the conclusion that it would be more appropriateand effective in reality to privilege more simple and prudent models than to complicatethings and generate additional errors and instability.
đang được dịch, vui lòng đợi..