Prudential Regulation Authority Rulebook

6.1

As per Step 5 of the PRA five-step framework, firms are expected to perform a recalculation of the MA in stressed conditions. This calculation should apply the same calculation method as that used to determine the MA for the purposes of calculating the TPs; including ensuring that the level of cash-flow matching is in line with the firm's matching tolerances.

6.2

Firms are also expected to validate the MA benefit assumed in the SCR calculation and the assumptions on which this MA benefit is based. There are a range of methodological approaches that firms could use to do this.

6.3

When validating the amount of MA assumed in the SCR calculation, the PRA expects firms to explicitly and separately cover:

• the assumed impact of the stress event on the level of MA benefit, and;
• the impact of any management actions taken to restore MA compliance, including the impact of these actions on the MA benefit.

6.4

For the purpose of ensuring the MA is appropriately reflected in firms’ internal models, the PRA expects firms to set out an approach that allows for the full range of risks and risk interactions to which the MA portfolio is exposed. Specifically for the MA portfolio, the PRA expects a firm to consider explicitly the impact of underwriting risk stresses the MA portfolio could suffer (eg a longevity stress and the associated requirement to find additional long-dated cash flows to match the additional long-dated liability cash flows arising after an adverse change in longevity assumptions).

6.5

Where a firm does not have an internal model covering all risks to which the MA portfolio is exposed, the PRA expects these risks and associated risk interactions to be considered in the own risk and solvency assessment (ORSA). Over time, firms should aim to incorporate all material risks and risk interactions to which the MA portfolio is exposed in their full or partial internal model, with consideration given as to any temporary measures that may be required.

6.6

Modelling management actions (to restore MA compliance) within the internal model is complex, especially for firms that determine their SCR using stochastic methods. Such actions can also change the ranking of scenarios and therefore the ultimate SCR. Firms should consider the effect of the management actions in a suitable number of scenarios across the probability distribution forecast to ensure an appropriate scenario ranking is achieved.

6.7

If a firm is using a variance-covariance approach to determine the SCR, the PRA expects the firm to consider whether any further adjustments are required to restore MA compliance in the SCR scenario. This can potentially be incorporated as an end-piece adjustment (eg via a non-linearity adjustment).

6.8

The PRA expects firms to validate the level of MA benefit assumed in the SCR calculation using a methodology that differs from the primary methodology used to calibrate the stressed MA. In particular, where a calibration method is highly reliant on expert judgement, the validation approach should aim to make use of historical data (if possible) to demonstrate the appropriateness of the output, and vice-versa. For the avoidance of doubt, the PRA does not expect firms to run two parallel models for the MA but the validation approaches used should be sufficiently independent from the core modelling methodology so as to give adequate comfort as to the appropriateness of the overall stressed MA and its underlying drivers.

6.9

Where firms’ models to determine the stressed FS are modifications of the approach used by EIOPA to determine the FS for the purposes of calculating the TPs, the PRA expects the models to be capable of replicating the FS provided by EIOPA in the same economic conditions. This should act as a check on whether the model is fit for purpose.

6.10

While the PRA recognises the importance of industry benchmarking surveys as a validation tool, care needs to be taken to ensure that comparable metrics are being assessed and that the surveys used take adequate account of potential differences between firms including the:

• approach to rebalancing; and
• concentrations of exposure.

6.11

Also, while simple metrics such as the MA in stress as a percentage of the spread widening have an intuitive appeal, the MA is a portfolio-level calculation and care needs to be taken to ensure that this is appropriately reflected in any conclusions based on benchmarking studies. Firms should not place material reliance on such an assessment for the purposes of validating the modelled matching adjustment in stress unless it is sure that the comparison is on a truly like-for-like basis.