Interpretation Note

[MAR21.71]

The relationship between the first and second sentences of [MAR21.71] is unclear. The description in the second sentence is not a good match for the correlation parameter in the first sentence, and it is not easy to determine the intention. So, Atoti FRTB supports three possible interpretations of [MAR21.71] and provides configuration parameters to choose between them.

The configuration parameters enable choosing between:

• Should $K_{other}$ be added after taking the square root in [MAR21.4](5) and [MAR21.5](4), or should the term $K_{other}^2$ be included inside the square root?
• Should the correlation with the “other” bucket ($\gamma_{b\ other}$) be 1 or 0?

By default, the configuration parameters are set so that the calculations match those in BCBS 352. An alternative, and equally valid, interpretation of [MAR21.71] would be to treat the “other” bucket the same as the other risk classes.

[MAR21.68]

For two tranches A and B in the same pool with attachment points A1 and B1, and detachment points A2 and B2, the overlap is calculated using MIN(A2 - A1, B2 - B1, MAX(A2 - B1, 0), MAX(B2 - A1, 0)) / MAX(A2 - A1, B2 - B1) and this is compared to the threshold of 0.8 (80%).

The Pool “N/A” is treated differently. Two different tranches in this pool will never be considered to overlap.