Erscheinung:03.12.2008 | Topic Own funds OpR Expert Group recommendation on the validation of the AMA (of 05.03.2008)
Preliminary remark
In its mandate the OpR Expert Group set itself the task of drawing up proposals for how the latitude that exists in the national implementation of the Basel and Brussels rules on operational risk might be utilised. The following Expert Group recommendation is a suggestion for the validation of the AMA. The recommendation is subject to its consistency with the decisions taken at the European level.
Please note:German version is binding
This translation is furnished for information purposes only and may refer to an older version of the text. The original German text is binding in all respects.
Sections 283-284 of the Solvency Regulation (as of 01.01.2007)
Section 283 - Audit
(1) 1The operational risk management processes and measurement systems shall be subject to regular reviews performed by internal or external auditors. 2These audits shall include both the relevant activities of the internal business units and those of the independent operational risk management function.
(2) Institutions shall ensure that data flows and processes associated with the operational risk measurement system are accessible to internal and external auditors without delay.
Section 284 - Quality of the measurement system
(1) 1Advanced Measurement Approaches shall use internal loss data, external data, scenario analysis and institution-specific business environment and internal control factors in calculating the capital requirement for operational risk. 2An institution shall combine these four elements appropriately in its Advanced Measurement Approach and document the fact. 3In particular, it shall ensure that in combining these elements, multiple counting of qualitative assessments or risk mitigation techniques is avoided.
(2) 1The capital requirement for operational risk calculated using an Advanced Measurement Approach shall include both expected and unexpected loss. 2Insofar as the institution determines the expected loss appropriately and proves that it takes a portion of the expected loss adequately into account in its internal business practices, BaFin will authorise it to reduce the capital requirement for operational risk by that portion of the expected loss.
(3) 1The Advanced Measurement Approach shall capture the major drivers of operational risk affecting the shape of the tail of the loss distribution. 2In particular, the capital requirement for operational risk shall capture potentially severe tail events, achieving a soundness standard comparable to a 99.9 % confidence interval over a one-year holding period.
(4) 1The institution shall use appropriate techniques in developing a model for measuring its operational risk and in validating the model. 2The validation processes, techniques and outcomes shall be documented.
Explanatory comments on section 284 (4)
Under the terms of the Directive, to approve AMA models the supervisory authority must assure itself that institutions' internal procedures for validating the risk measurement system are satisfactory. Consequently, institutions must have appropriate internal validation techniques.
The internal validation of the model and the techniques used for this purpose are the responsibility of the institution. The objective of the validation is to ensure the quality of the measurement system. The validation techniques must include an adequate analysis of the measurement system. The validation techniques must be documented and included in the audits conducted by internal or external auditors. The model input, the model assumptions and the model output have to be validated. The validation of the measurement system should include qualitative and quantitative elements, since an adequate mathematical/statistical validation of the whole model or individual components of models is frequently not yet possible.
Internal data, external data, scenario analyses and business environment and internal control factors have to be incorporated in the measurement system. The elements included in the model have to be examined whether they adequately represent the institution's risk profile. The data used in the risk measurement system and the assumptions made must be realistic and sufficiently accurate. The validation of this aspect should be repeated at appropriate intervals in order to incorporate new elements as well.
The modelling methodology must be appropriate for the size and complexity of the institution. Any change in the institution's risk that is reflected in the data and assumptions used in the model must lead to a coherent change in the model's results. Also if loss severities and frequencies that significantly exceed the institution’s data are used as model input, the model output must indicate a coherent change in risk. This is to be demonstrated by sensitivity analyses. The measurement system will need to be re-validated as it is enhanced and in the event of material changes in the institution, its activities or its business environment.
The results produced by the model may be validated with the aid of e. g. comparisons between business lines and, if necessary, comparisons with competitors operating in the same market environment and, in the case of a loss-data based AMA, with scenario tests or if possible by statistical techniques. Furthermore, the results should be examined to see whether changes in risk shown by the instruments used for controlling operational risk are consistent with the changes in the elements included in the model and also, in the medium term, with the results produced by the model. A comparison with the capital requirement that would be produced in the Basic Indicator or Standardised Approach may also be made, but is not sufficient for validation purposes.
