Solvency UK Matching Adjustment reforms: What about the Internal Model?

At a glance

In our previous blog¹ on the CP19/23 consultation, we explored how the PRA’s proposals are expected to impact the Matching Adjustment (MA) used in the calculation of life insurers’ Technical Provisions for annuity business. These changes will also have knock-on implications for the Internal Model used by these firms to calculate the Solvency Capital Requirement (SCR).

In this blog, we discuss these implications and set out the key areas for firms to consider as they change their Internal Models to allow for the proposals in CP19/23. These key areas are:

• Notched credit ratings. The Solvency UK reforms require firms to use more granular ‘notched’ asset credit ratings in the Fundamental Spread (FS) calculation for the base MA, but firms can choose whether to use this notching in their Internal Model calibrations. Where firms opt not to apply notching in the SCR calculation, they will be expected to justify such an approach.
• Removal of the sub-investment grade (SIG) cap. The PRA’s proposed removal of the SIG cap is likely to increase the MA benefit earned on SIG investments. Firms will need to review their SIG modelling capabilities and the relative materiality of the calibration for SIG assets before removing the cap within the Internal Model. This review may lead to changes in SIG calibrations.
• FS add-ons. The Solvency UK reforms require firms to apply FS add-ons in the base MA calculation where they consider that the risks retained by the firm are not sufficiently captured by the basic FS calibration. Where FS add-ons are applied for a given asset type, firms will need to consider how any additional risks captured by the add-on behave under stressed conditions.
• Assets with highly predictable cash flows (HP assets). The Solvency UK reforms allow for the inclusion of assets with “highly predictable” cash flows in MA portfolios, subject to certain conditions. This potentially offers firms a wider universe of assets in which to invest, but firms will need to ensure that any new risks are adequately modelled within Internal Models.
• MA breaches. The PRA has proposed a less penal treatment of breaches of MA requirements, which may impact firms’ current modelling of management actions under stress in Internal Model calibrations.
• Matching tests. The PRA has introduced two new cash flow matching tests for firms with HP assets in their portfolios and minor amendments to existing matching tests. Whilst applicable in base conditions, these will also increase the matching requirements under stress for Internal Model firms.
• Data limitations. In adopting notched ratings, modelling HP assets and FS add-ons, or reviewing SIG approaches, firms may need to consider new, potentially limited datasets to inform the Internal Model calibrations. Firms may need to exercise expert judgement where there is a lack of credible or high-quality data to appropriately calibrate stressed scenarios, as well as ensure that the lack of data does not undermine compliance with the Prudent Person Principle (PPP).
• Interaction with CP12/23 Internal Model proposals. The CP12/23 consultation proposes a more flexible, permissions-based Internal Model application process, which allows the PRA greater discretion but also a wider range of tools to support this. Firms may consider accepting new safeguards such as capital add-ons (CAOs) to recognise any residual model limitations (RMLs) in the Internal Model in the short-term (e.g. YE24), with a clear view of the developments required down the line to address such limitations.

Given the short timelines available to make model changes before YE24, firms will need to quickly form a view on their approaches to each of the areas described above. Firms should identify changes needed for the end of the year and those that could be made later and prioritise accordingly. This will help minimise nasty surprises close to year-end and provide valuable insight on the net effect on capital and solvency of the overall reforms.

Who this blog is for

Board members, senior executives, and actuaries of UK insurers with MA and Internal Model approvals who work on balance sheet management, pricing, reporting, capital optimisation, risk, finance, and compliance.

Implications of the Solvency UK Matching Adjustment reforms for Internal Model firms

The PRA’s proposed changes to the Matching Adjustment (MA) as part of the Solvency UK reforms, outlined in the consultation paper CP19/23, will impact the Internal Models used by firms to calculate the Solvency Capital Requirement (SCR) for their annuity business.

Previously, we explored¹ how the PRA’s proposals are expected to impact the calculation of and ongoing compliance around the MA used in the base balance sheet. In this blog, we discuss the main implications of these proposals for Internal Models. These are relevant for firms with both MA and Internal Model approvals, including all the current UK bulk purchase annuity (BPA) providers and firms that are planning to enter the market².

Below, we work through the key areas firms need to consider as they alter their Internal Models to allow for the knock-on impacts from CP19/23.

A. Notched credit ratings

Currently, the FS is calibrated and published at a letter credit rating level (AAA, AA, A, BBB, etc.). The Solvency UK reforms introduce a requirement for firms to use more granular ‘notched’ credit ratings in the base FS calculation.

Firms are not explicitly required to calibrate their Internal Model stresses using notched ratings, but those that choose not to would need to be able to justify differences in the granularity of ratings in their base MA and SCR calculations. Notably, among other considerations, firms may need to demonstrate that there is no significant bias in the asset mix towards the lowest notch at each credit rating. Some firms will have had investment limits in place to mitigate such a bias during the investment decision stage, but other firms may have not considered this and therefore may find it more difficult to justify not including notched ratings within the Internal Model.

Firms that opt to introduce notched ratings in Internal Models should consider whether this will increase the sensitivity of the Internal Model to modelled downgrades and therefore increase the SCR.

B. Removal of the SIG cap

Under the current Solvency II regime, the MA benefit earned on SIG assets is not permitted to exceed the MA for BBB-rated assets of the same duration and asset class. The Solvency UK reforms remove this requirement.

This will improve the economic attractiveness of holding SIG assets, which could be particularly relevant for firms investing in illiquid assets, where there tends to be a larger concentration of SIG assets. Whether or not firms do increase their holdings in SIG assets will ultimately depend on their risk appetite, strategic asset allocation objectives and competitiveness considerations.

Removal of the SIG cap within the Internal Model would be expected to provide a reduction in capital, even if firms have minimal holdings of SIG assets. This would be driven by downgrades into SIG ratings in stress being less penal than before.

If firms intend to remove the SIG cap within their Internal Models, they should consider whether the SIG stresses in their Internal Models would remain appropriate or if more robust calibrations will need to be introduced. In particular, firms should consider the following questions:

• Were simplified calibrations or models used for SIG assets, on the basis that the SIG cap would bite and preclude the need for a more detailed calibration?
• Do the SIG calibrations capture all the nuances of SIG assets, as discussed in CP19/23?
• Do assumed management actions under stress (for example, selling off SIG assets if they downgrade) need to be reconsidered?
• Does the higher materiality from any increase in SIG assets in the base position necessitate more accurate modelling?

These considerations may suggest a need to develop the Internal Model further, which may take some time. If it is not possible or desirable to deliver these changes by the end of 2024, an option for firms is to keep the SIG cap in place, accepting that this would mean no reduction in SCR, until further development can be completed.

C. FS add-ons

The PRA proposes that firms consider voluntary FS add-ons for assets where the basic FS parameters (based on corporate and government bonds) do not capture all risks retained from holding those assets. The PRA requires firms to consider how these FS add-ons may behave in stress, noting in paragraph 6.13 of the CP that it “would not expect any voluntary FS addition to automatically result in a reduction in the SCR… consider the extent to which the risks that are allowed for in a voluntary FS addition are also allowed for within the calibrations underlying the SCR and consequently how they may change in stress conditions.”

If companies hold assets with non-zero FS add-ons, this requirement could be met by either:

• Directly modelling the FS add-on under stress, leading to either no change to the SCR where the FS add-on remains constant under stress or a change to the SCR where the FS add-on changes under stress; or
• Maintaining the existing stress approach and explaining how the risks addressed by an FS add-on in the base MA calibration are already reflected within the Internal Model calibrations under stressed conditions. In this situation, the stressed FS would remain unchanged, with the difference between the base and stressed positions decreasing, leading to a lower SCR. Given the PRA’s note on this, strong justification would be expected for such an approach.

Regardless of how this is allowed for, firms will need a good understanding of the retained risk elements under both base and stressed conditions to justify the approach. One option is for firms to decompose any FS add-on into the sources of retained risk and articulate how such risks behave under stress. For this purpose, firms might choose to leverage analysis performed on each component of the base FS as part of the MA attestation requirement, but this will need to be subject to robust validation to ensure that the data is fit for purpose under both base and stressed conditions.

Where applicable, firms will also need to consider how the modelling of the FS add-on under stress interacts with the prescribed minimum FS add-on for HP assets of 10bps.

D. HP assets

The Solvency UK reforms aim to enhance investment flexibility by making HP assets MA-eligible, provided that the MA benefit for such assets is no more than 10% of the total MA benefit and the cash flows are contractually bound.

Where firms intend to use this increased flexibility to invest in new asset classes, they will need to consider their ability to model the risks underlying these assets under stress and develop their Internal Models to reflect them. This may require time to develop, with Internal Model changes only feeding through during 2025 or later.

There is a risk that firms may be compelled to reclassify some of their current assets as HP assets over 2024. Where this is the case, focus will be required on the modelling of these assets, including within the Internal Model, in order to meet YE24 deadlines.

Investment in HP assets may give some insurers access to new asset types whose investment returns may be less correlated to those of typical market instruments used to back MA liabilities, such as corporate bonds. Thus, firms moving into this space could consider whether it is appropriate to allow for diversification benefits within the Internal Model SCR calculations between HP assets and assets with fixed cash flows.

E. MA breaches

Under the Solvency II regime, firms’ MA approvals are revoked if breaches are not resolved within two months. The Solvency UK reforms reduce the severity of breaches such that the MA benefit is gradually reduced instead, with the reduction being 10% of the MA, increasing by an additional 10% for each month after the two-month window.

This is likely to have an impact on the range of management actions available to Internal Model firms under stress. For example, in the case of rating downgrade or default of an asset within the MA portfolio, firms may have greater flexibility in terms of assuming that appropriate replacement assets can be sourced within a longer time frame than two months, at the expense of a temporary decrease in the overall MA benefit.

F. Matching tests

The PRA proposes two new cash flow matching tests for firms using HP assets in their MA portfolios, as well as minor amendments to the existing Tests 1 and 2. The two new tests aim to ensure that firms’ exposure to reinvestment and liquidity risk introduced by HP assets is sufficiently low, and the amendment to Test 2 requires that HP cash flows are modelled consistently with the relevant stress scenarios. Firms will need to update Internal Models to be able to perform the new or amended matching tests. These tests will need to be satisfied under stress and will provide further considerations for firms when modelling stress outcomes.

G. Data limitations

A recurring theme across all the items mentioned above is the availability of data to calibrate Internal Model stresses. For example, firms that decide to adopt notched ratings in their Internal Models should consider the availability and credibility of data at a notched rating level during historic periods of economic stress. Default and transition rates during the most significant historical downturn – the 1930s – may not be available at such a granular level. Firms may consider fitting a curve to their current stresses at a letter rating level to infer notched stresses, but this will require additional judgement and may take time to build into models.

Similarly, many HP and SIG assets also have limited experience data available. Where this is the case, it may be useful to leverage default and downgrade data from external credit assessment institutions and recognise the need for significant expert judgement to appropriately derive a stressed calibration. The PRA’s proposal that firms must be able to evidence how their MA portfolios comply with the PPP will increase the need for firms to showcase how they can identify, measure, and manage the risks on these assets, despite the challenges around data limitations.

H. Interaction with CP12/23 Internal Model proposals

The PRA’s less binary, permissions-based approach to Internal Model approvals, proposed in its first consultation paper CP12/23³, may allow firms the flexibility of recognising any modelling deficiencies resulting from the data challenges discussed above as RMLs, at the cost of holding PRA-enforced CAOs or model limitation adjustments (MLAs) while these are resolved.

Although firms will hope to avoid such safeguards, given the speed with which the MA changes will come into force, we may see some use of these across the industry for at least FY24. It would therefore be sensible for firms to ensure that there is a clear development plan to address any modelling limitations underlying a CAO or MLA in a timely manner.

Conclusion

Internal Model firms will need to assess the potential impact of the Solvency UK reforms to the MA on their stressed calibrations and consider how these may interact with the proposed changes to Internal Model regulations detailed in the PRA’s CP12/23 consultation paper. There are several potential challenges arising from data availability, and firms should consider how they will mitigate these to ensure that their Internal Model calibrations remain appropriate.

Firms should quickly ascertain which changes should be made ahead of YE24 and identify where CAOs or MLAs are likely to be required due to modelling limitations and what the longer-term Internal Model development plan is. Early consideration of this will help minimise nasty surprises close to year-end and provide valuable insight on the net effect on capital and solvency of the overall reforms.

__________________________________________________________________________________________

References

¹ Solvency UK Matching Adjustment: Another piece of the puzzle falls into place | Deloitte UK

² Breaking into the BPA market: Challenges for new entrants | Deloitte UK

³ Solvency UK: Moving closer to reality | Deloitte UK