Читать книгу Non-financial Risk Management in the Financial Industry - Группа авторов - Страница 113
Table 6: Example of KRI threshold calibration and fine-tuning
Оглавление| 1 – Large EU Banking Group The bank needed to set the limit, caution and target thresholds for the following KRI: The bank firstly determined its data series collecting data points from all legal entities in the group and adding them to the data series of the previous two years. Then two different distributions were carved out of the original one: the first one referred to large company customers, the second one to small-size company customers. In accordance with the overall RAF methodology, the bank set the following thresholds: Target: values in 80th–90th percentile. Caution: values above the 95th but below the 99th percentile. Limit: value above the 99th percentile. This setting was backed up by simulations and backtesting, which showed that a breach would have occurred only three times in the past three years, thus being consistent with the definition of extreme values. Furthermore, a benchmark analysis was run, and it confirmed that the main market peers, comparable for volume of transactions and geographical fragmentation of business, used the same thresholds on average. |
