DP2/22 – Potential Reforms to Risk Margin and Matching Adjustment within Solvency II

Executive summary

The Prudential Regulation Authority (PRA) notes the publication by HM Treasury (HMT) of its consultation on the Solvency II Review (the Review).

The PRA supports the objectives of the Review and continues to work closely with HMT on the potential reforms. This Discussion Paper (DP) sets out the PRA’s current views on some key aspects of the potential reform package.

The PRA's approach is grounded by its statutory objectives given by Parliament for safety and soundness, policyholder protection, and the secondary competition objective. The PRA also has regard to the range of matters it is obliged to consider when making policy, including the impact on sustainable growth, innovation, competitiveness, trade, and climate change. Many of these matters align closely with the Government’s wider objectives for the Review.^{footnote [1]}

HMT’s consultation considers its proposed reforms to Solvency II could result in a release of possibly as much as 10% or even 15% of the capital held by life insurers. It proposes reducing the risk margin for long term life insurers by 60-70%, and affirms that HMT is considering the case for reform to the fundamental spread (FS) used to calculate the matching adjustment (MA). It notes the link between the FS and other parts of its proposed reforms, including to the risk margin and the eligibility requirements for assets in insurers’ MA portfolios. The MA matters are primarily of relevance to the annuity sector.

In working with HMT, the PRA has explained its view on which potential combinations of reforms to the FS and the risk margin could be consistent with its statutory objectives, and which would not. The PRA has drawn on available evidence, its engagement with insurers and its assessment of the estimated reduction in aggregate capital levels for the insurance sector (and therefore safety and soundness and policyholder protection) that the reform options would imply. The data gathered in the 2021 Quantitative Impact Study (QIS) has provided useful insights in informing the PRA’s view.^{footnote [2]}

Provided that satisfactory reforms to the FS and risk margin are achieved, the PRA considers there exist packages within the indicative ranges included in HMT’s consultation that would be consistent with its statutory objectives and that would also achieve the broader objectives of the review around competitiveness and long-term investment.

Objectives of the review

The PRA considers that reforms need to ensure the long-term safety and soundness of the UK insurance industry and deliver an appropriate degree of policyholder protection. UK insurers need to maintain the financial strength on which their policyholders depend to have confidence that claims will be paid, and for the security of their pension income and financial wellbeing in retirement. In practice, the adequacy of capital for insurers is calculated over a one-year time horizon, with the key test being that an insurer is able to transfer its business to another if it is in distress. The risk margin and MA are key to the calculation of this transfer value.

Maintaining the confidence of policyholders and investors is essential if the insurance sector is to remain strong and competitive and to play its part as an important provider of long-term investment in the UK economy. The PRA considers that the reforms outlined below would support UK insurers’ ability to invest for long-term growth, and facilitate a thriving, competitive and safe UK insurance sector.

The case for reform

The trend in transfers from company pension schemes to life insurers over the last five years is projected to continue, with rapid growth in the sector. At the same time the assets insurers choose to back these liabilities are changing, and Solvency II’s MA treatment was not explicitly designed to cover these new assets. The package of reforms is an opportunity to address more appropriately the risks life insurers are taking.

The current design of the FS means there is a risk insurers recognise profits upfront on their investments in MA portfolios that may not actually be realised in the future. The current prudence in the risk margin gives some mitigation of this risk. So it is important that an overall package of reforms which includes reducing the risk margin also includes a strengthening of the FS in order to maintain insurers’ safety and soundness and protect policyholders, consistent with the PRA’s statutory objectives.

1. Introduction

1. This DP sets out the PRA’s current assessment of overall reform outcomes arising from key aspects of HMT’s review of Solvency II – particularly reforms to the FS and risk margin. The content of this DP is primarily of relevance to the annuity sector. It will also be of interest to other insurance and reinsurance undertakings and industry stakeholders.

The Solvency II Review

2. The Solvency II regime came into force in the UK on Friday 1 January 2016. Prior to the UK leaving the European Union (EU) on Thursday 31 December 2020, the PRA had been contributing to the EU’s five-year review of Solvency II. With the UK leaving the EU, this has provided an important opportunity to deal with some key areas of Solvency II that can be improved, while maintaining a regime that is robust in protecting policyholders, is well-respected internationally, and that befits the UK’s position as one of the leading international insurance centres. In particular, there is an opportunity to tailor the regime so that it fits the UK market better, and is more efficient and coherent.

3. In June 2020, the Government announced a review of the Solvency II framework in the UK, and in October 2020 it launched a ‘Call for Evidence’^{footnote [3]} that set out the areas in scope for this review. The information provided by respondents to the Call for Evidence was subsequently shared with the PRA. Drawing on this information, the PRA designed and launched the QIS in July 2021 in order to gather data from firms to model the impact of potential reforms to the risk margin and MA. This DP focuses on the interaction of reforms to these two key areas of the Review.

4. There are other planned reforms^{footnote [4]} to Solvency II covered in both the Call for Evidence and HMT’s consultation that are not considered directly in this paper. In addition, the PRA is aware that any reforms to the FS may interact with the quantification of the Solvency Capital Requirement (SCR); this issue is discussed briefly in the technical annex to this DP, ‘Solvency II Review: Matching Adjustment and reforms to the fundamental spread’ (the technical annex).

The reforms as a package

5. The PRA is of the view that decisions on changes to the risk margin need to be taken together with decisions on the FS when assessing the overall impact of reforms on its statutory objectives. When considering the potential calibration of these reforms, the PRA has also taken into account whether long-term life insurers would continue to hold sufficient assets to be able to transfer their liabilities to a third party in the event of their failure (ie whether the UK would continue to operate a ‘going concern’ prudential regime in line with international standards). It is on this basis that, even for long-term business, an insurer’s SCR is assessed at a 99.5% level of confidence over one year.

6. This DP provides the PRA’s current view of potential reform outcomes focusing on changes to the risk margin and FS, and the estimated impact on overall capital levels and the valuation of technical provisions. Changes to the risk margin and FS will have the most direct impact on the financial position of insurers and therefore their safety and soundness.

Discussion paper structure

7. This DP is structured as follows:

• Chapter 2 describes the PRA’s current position on reform of the MA;
• Chapter 3 describes the PRA’s current position on reform of the risk margin;
• Chapter 4 describes the PRA’s considerations on validation of potential reform packages;
• Chapter 5 describes the expected impact of the reforms outlined in this DP; and
• Technical Annex: Solvency II Review: Matching Adjustment and reforms to the fundamental spread.

Responses and next steps

8. Any reforms from the Review are likely to require changes in due course to domestic legislation (which is the responsibility of HMT and Parliament), and also changes to PRA rules and supervisory expectations. HMT has explained it will be considering feedback to its consultation before deciding which reforms should be implemented through changes to legislation, and which might be taken forward by the PRA. The PRA will continue to work closely with HMT on the Review, so that it is in a position to consult as soon as possible on any proposed changes to its rules and supervisory expectations. When preparing its future consultation, the PRA will perform its usual assessment of the costs and benefits of any proposed changes against its statutory objectives and against relevant ‘have regards’.

9. To inform this further work, the PRA welcomes feedback on the matters discussed in this DP, in the technical annex on the MA and FS, and more broadly on the reform package. In particular, we would welcome views on:

• The potential formulation for the Credit Risk Premium as outlined in chapter 2;
• The design and calibration of the risk margin as described in chapter 3; and
• The extent to which combined potential reforms to the risk margin and fundamental spread result in a package that meets the PRA’s objectives.

The PRA notes that the matters discussed in this DP are largely technical in nature, and is seeking feedback on these areas to supplement the answers provided to HMT’s consultation.

10. This DP closes on Thursday 21 July 2022. The PRA invites comments on this analysis to help develop its thinking and to inform any future PRA consultation proposals. Please address any comments or enquiries to Solvency2Review@bankofengland.co.uk. Your responses may be shared with HMT. This means HMT may review the responses and may also contact you to clarify aspects of your response.

11. The PRA will also be setting up further opportunities for industry engagement to discuss these issues in more detail.

2. Reform of the Matching Adjustment

12. The MA is a mechanism that allows insurers to recognise upfront as capital resources a proportion of the spread (in excess of the risk-free rate) they project to earn over the future lifetime on the assets matching their MA liabilities. The PRA considers that inclusion of the MA in the solvency framework is justified since (i) spreads on assets exposed to credit risk are generally materially greater than the level required to compensate for the underlying credit risk, and so include a significant ‘liquidity premium’;^{footnote [5]} and (ii) insurers with predictable liability cash flows that are closely matched by cash flows from a portfolio of MA-eligible assets are not materially exposed to liquidity risk, and so can reasonably recognise the liquidity premium upfront as capital.

13. The MA provides a strong incentive for life insurance firms to match their asset and liability cash flows, which reduces prudential risks. However, allowing firms to recognise profits before they are earned is a valuable benefit which needs to be calibrated carefully to avoid the risk that insurers recognise returns upfront which may not be earned in future.

14. The calibration of the MA requires balance. It should not be calibrated such that the level of policyholder security is excessive in relation to the risks faced by insurers in respect of their MA portfolios. An appropriate MA treatment promotes safety and soundness of firms and policyholder protection, results in a competitive annuity market, and incentivises firms to invest in a wide range of long-term, illiquid, fixed-interest assets which can also provide wider economic benefits. Firms’ management of annuity business, as supported by the MA, is therefore aligned with the PRA’s secondary competition objective and ‘have regards’ in relation to proportionality, competitiveness, innovation and long-term productive investment.

15. However, the MA is a particularly material benefit for insurers writing annuity business. The appropriateness of its design and calibration is therefore important, noting in particular:

• the importance of annuities for the retirement provision for policyholders;
• the long-term nature of the risks to which annuity writers are exposed; and
• the difficulty of de-risking the asset portfolio once written (since it is likely to be difficult to trade out of relatively illiquid assets while also maintaining the MA that has already been reflected in a firm’s solvency capital).

16. It is therefore particularly important to maintain an appropriate degree of policyholder security for annuity business, including the ability to transfer the business to another insurer if necessary. Along with the risk margin, the MA is key to the determination of the liabilities as a ‘transfer value’.

The Fundamental Spread

17. The FS is the allowance made for the risks that insurers are assumed to retain, when calculating the MA benefit.

18. The PRA considers that firms’ MA portfolio assets are exposed to two sources of retained credit risk: (i) expected loss due to default (EL); and (ii) uncertainty around that EL for which a willing arm’s length third party would demand a premium for taking on the risk. This second element can be referred to as the credit risk premium (CRP). The PRA is concerned that issues with certain aspects of the current FS design mean it is not appropriately allowing for both of these elements.

19. More specifically, the PRA has three concerns with the current FS construct:

1. The FS does not capture all retained risks which insurers face and as such its level (in basis points) is generally too low. This is because the FS does not fully and explicitly allow for uncertainty around credit risk (the CRP). The PRA considers that cash-flow matched insurers remain fully exposed to credit risk (in relation to both defaults and downgrades) and that it is unsound to recognise upfront as capital the compensation for credit risk that they hope to earn over the lifetime of their assets.
2. The FS is not sensitive to differences in risks across asset classes for a given currency, sector and Credit Quality Step (CQS). It assumes that all assets of the same currency, sector, term and CQS have the same amount of credit risk, calibrated to historical externally rated corporate and government bond data. This means firms are incentivised to hold high spread-for-rating assets, as all of the excess spread is taken credit for as MA upfront, so creating Tier 1 capital.
3. The FS does not adjust to reflect structural shifts in the credit environment over time, unless there are actual defaults or downgrades. This means that any potential signal of change in credit risk contained in credit spread movements is discarded. The PRA recognises that spreads can be more volatile than is justified by the genuine credit risk to future cash flows, and that significant smoothing is desirable to prevent unwarranted balance sheet volatility. However, insofar as changes in spreads signal genuine changes in credit risk, the PRA considers that it is inappropriate for insurers’ balance sheets to entirely disregard such signals.

20. These issues with the FS mean there is a risk that the MA benefit currently being taken by firms is too high, particularly as insurers’ investments have changed over time and as the proportion of investment assets rated and valued by insurers themselves has increased. This risk becomes more acute as action is taken to remove excessive prudence from the risk margin. Although the PRA can seek to address some of these limitations through its supervision of individual firms, this is not an effective substitute for a properly constructed FS treatment for the insurance sector as a whole.

21. The technical annex details some of the PRA’s findings regarding current asset holdings within MA portfolios, evidencing:

• the disconnect between the assessment of risks retained and market spreads, even to the extent of MA benefit being greater than the market value of the asset in some cases;
• the clear incentive to seek out those assets with highest possible credit spread within each rating category; and
• the skewed incentives to pursue certain asset types over others, not obviously aligned with the asset’s risk profile, or the extent to which it provides benefits to the wider economy.

22. The PRA considers that reform to the FS is needed to allow appropriately for the risks retained, so that insurers do not recognise excess levels of profit upfront. An appropriate FS will ensure that the compensation for credit risk is recognised only as the credit risk being borne by the insurer runs off. Making proper allowance for credit risk will increase firms’ ability to continue to hold productive assets through changing credit conditions over their lifetime. Further, this will level the MA benefit playing field which currently is skewed not towards productive assets but in favour of assets that provide the highest immediate MA benefit because they have a high spread relative to rating.

23. This reform will then facilitate other reforms within the Review, including to the risk margin and reforms aimed at widening the range of assets eligible for the MA treatment and streamlining the approval process, which can support the Government's objective to increase investment in long-term productive finance.

24. To reflect adequately the risks faced by insurers, the PRA considers that the FS should include a CRP. There are various ways this could be achieved in practice, but the PRA currently considers that a CRP needs to be calibrated to deliver an outcome equivalent to at least 35% of credit spreads on average through the cycle. This judgement has been informed by a review of academic research (which the PRA judges supports a range of 35% to 55%, with 35% at the lower end of the range of estimates for a CRP), experience of regulating the insurance sector during the 2007-09 financial crisis and under the Solvency II regime, and supporting analysis including stress testing against historic experience. Further detail on the evidence that supports this judgement has been set out in the technical annex.

25. In addition, the PRA considers that reforms to the FS should be implemented in a way that avoids undue volatility of life insurers’ balance sheets. HMT has outlined in its consultation document one option for reform to the FS, including a possible design of a CRP component. Under this approach, the CRP for an asset could be set as a portion of the average spread on a relevant index over a chosen period, plus a portion of the difference between the asset’s current spread and that of the index (an ‘index-spread’ approach). The portions and the averaging period would be chosen so as to reflect credit risk appropriately, while avoiding undue volatility. The CRP or FS would also be expected to include other through the cycle features, such as caps and floors, as explored in the technical annex. The technical annex also contains more detail on the overall approach and how it could be implemented, which will allow further information-gathering and stakeholder engagement.

26. In developing its views on reform of the FS, and on reforms to the MA asset eligibility criteria and approval process, the PRA will also continue to have regard to the potential impact on insurers’ investments and how these may interact with sustainable growth, innovation, competitiveness, trade and climate change. These factors are well-aligned with the wider objectives of the Review.

27. The PRA welcomes feedback on the matters outlined here, and in the accompanying technical annex, including the potential formulation for the CRP.

3. Reform of the Risk Margin

28. The risk margin is an amount added to the best estimate liabilities (BEL) so that the total technical provisions represent a transfer value – the amount for which the liabilities could be transferred to a willing third party. This gives a high degree of confidence that, if a firm gets into difficulty, its insurance liabilities can be transferred to another viable firm. In particular, the ability to transfer the business is key for providing a bridge between the one year 99.5% confidence level underpinning the SCR, and the longer-term risks inherent in insurance business.

29. The risk margin is currently calculated using a ‘cost-of-capital’ approach. It is computed as the cost of the regulatory capital that would be needed to support the liabilities over their lifetime.

30. The PRA agrees that the risk margin should be reformed to deal with concerns that it is too sensitive to movements in interest rates and too high when interest rates are low. Responses to the Call for Evidence, and the PRA’s analysis of information received through the QIS, both suggest that the issues with the current design are much less acute for non-life than for life business. This is because of differences in the nature of the underlying risks, and more specifically the typical duration of the liabilities.

31. To support risk margin reform, the QIS collected data from insurance firms that has allowed PRA to model a range of possible designs and calibrations. The two approaches explored were: (i) a modification to the existing cost-of-capital approach, and (ii) a percentile approach similar to that used by the International Association of Insurance Supervisors (IAIS) for the Margin Over Current Estimate (MOCE), the equivalent to the risk margin within the Insurance Capital Standard (ICS). These two approaches can be calibrated such that they give similar outcomes at an aggregate level, but given the differences in their designs, they can have very different impacts at the individual firm level. The two approaches also produce different outcomes across different economic conditions; namely, the percentile approach is comparatively less sensitive to interest-rate movements, but in contrast to the existing cost-of-capital approach, it is spread-sensitive.

32. The PRA considers that modifying the existing cost-of-capital approach has the greater advantages, and so currently prefers this method for both life and non-life insurers, for the reasons set out in HMT’s consultation.^{footnote [6]} The PRA’s assessment has taken into account evidence from a number of sources.^{footnote [7]} Whichever method is used to implement risk margin reform, the PRA will ensure that UK Internationally Active Insurance Groups are held to at least international minimum standards.

33. Under the modified cost-of-capital approach, the cost-of-capital rate used in the formula could be amended. In addition, a new tapering parameter, lambda, could be introduced to allow progressively lower weight to be given to each year of projected future capital requirements.^{footnote [8]}

34. The PRA has considered the different arguments that have been put forward for introducing a tapering parameter lambda with a value of less than one. The PRA’s view is that the strongest of these arguments predominantly apply to life insurance risks, and in contrast for some non-life risks there are arguments that would suggest lambda should not be lower than one. For these reasons, and also considering that the problems with the current risk margin design are most acute for life business, the PRA’s view is that a tapering parameter should only be applied to life insurance business.^{footnote [9]}

35. The PRA has also considered the arguments for lowering the cost-of-capital rate from its current level of 6%. In the PRA’s view, there is a case for this, with the arguments applying equally to both life and non-life business.

36. However, there are limits to the calibrations of both the lambda parameter and the cost-of-capital rate that can be supported by available evidence.

37. Modifying the existing cost-of-capital approach to the edge of what the PRA’s technical analysis supports could result in a reduction in the risk margin of around 60% for long-term life business under current economic conditions. The PRA considers that a risk margin at this level, when combined with an FS including a CRP calibrated to achieve an outcome equivalent to 35% of credit spreads on average through the cycle, falls within the range that plausibly results in technical provisions equivalent to a ‘transfer value’ (see Chapter 4 below) and so is consistent with the PRA’s policyholder protection objective.

38. For non-life insurance business, the PRA considers that a deep cut to the risk margin is not justified, for the reasons outlined above - ie responses to the Call for Evidence and QIS suggested that the issues with the current design are much less acute for non-life than for life business. However, a plausible technical justification can be made for changes that would reduce the risk margin by around 30%. A reduction of this size would also support the review’s objective to spur an internationally competitive insurance sector, not least as the PRA estimates that a cut of around 30% for non-life firms would be similar to the effect of the European Commission’s (EC) proposed reforms.

4. Validating reform packages against observed transfer values for long-term life business

39. To validate whether the technical provisions for long-term life insurance genuinely reflect a market value, it is desirable to compare them against actual data on transfers of insurance business. Transfers of insurance business are relatively infrequent and their pricing reflects many idiosyncratic factors. However, several respondents to the Call for Evidence referenced the active market for transferring longevity risk, and commented that prices in this market suggested the current risk margin was too high particularly for annuity business.

40. A small number of respondents to the Call for Evidence included estimates of the typical premium for longevity risk transfers. These estimates fell in the range of 3-6% of the BEL, with the modal response being 5%. The PRA understands from subsequent discussions with firms that the full range of premiums is far wider than this, with the lower end mostly applicable to straightforward, shorter-term deals (which can also transact at prices lower than 3%), while more complex deals, or those involving a significant proportion of deferred annuitants, can transact at prices much higher than 6%.^{footnote [10]}

41. The PRA agrees that pricing data from the longevity risk transfer market is a useful input to help validate the potential outcome of risk margin reform, particularly for annuity business.^{footnote [11]} One important caveat is that although longevity risk is the dominant risk for annuity business, it is not the only risk. A transfer where the acquirer accepts other risks such as counterparty, expense and operational risk would likely require a higher premium than would be charged for accepting the longevity risk in isolation.

42. It is also worth noting that the PRA understands that current longevity risk transfers often involve only the most ‘attractive’ of the longevity risk on a firm’s books, so transfer of all longevity risk for a given portfolio (ie including any residual idiosyncratic longevity risk) would likely require a higher premium.

43. If the risk margin were reduced by 60% or more in isolation, without any change to the current FS, then the PRA’s analysis suggests that technical provisions for annuity business would be valued at levels below those implied by the evidence on typical longevity transfer prices. This brings a risk of disorderly failure – a firm in difficulty could find itself unable to secure a transfer to a viable third party, absent an injection of further funds. This would undermine the objectives of safety and soundness and policyholder protection, as well as the concept of Solvency II as a ‘going concern’ regime.

44. In contrast, a reduction of around 60% to the risk margin, combined with FS reform to include a CRP equivalent to 35% of credit spreads through the cycle, would not be inconsistent with this transfer value evidence. Such a package would result in annuity technical provisions that, in the PRA’s estimation, would fall within the range that would plausibly result in technical provisions equivalent to a transfer value, and so be consistent with the PRA’s safety and soundness and policyholder protection objectives.

5. Impact of reforms on overall capital levels, safety and soundness and investments

45. The PRA has estimated the potential capital impact of these reforms based on data gathered from the QIS. In this DP, ‘percentage change in capital’ is calculated by estimating the reduction in regulatory requirements from the reforms, and expressing them as a percentage of current life industry Own Funds. All calculations are performed assuming the reforms are fully phased in, and any existing transitionals have run off in full. Capital release figures are given under economic conditions similar to those at year-end 2020, the date used for the QIS data collection. PRA analysis suggests that capital release figures will be similar under economic conditions similar to those at year-end 2021.

46. The precise impact will vary as proposed calibrations are finalised, but the PRA currently assesses that a package of a c.60% risk margin reduction (for life business) and an FS calibrated to include a CRP equivalent to 35% of credit spreads over the cycle would release between 10% and 15% of capital from the life sector in current economic conditions. A package which has a CRP of 25% (significantly below the level the PRA considers appropriate) would be likely to release between 15% and 20% of capital, and a package with a CRP of 45% would be likely to release between 5% and 10% of capital. Cutting the Risk Margin by around 70%, as opposed to around 60%, would release 2%-3% more capital.

47. Insurers would be free to choose how to use any capital released from these reforms. If they choose to use it to support the writing of new business, the PRA’s preliminary assessment is that a package of a c.60% risk margin reduction and an FS with a 35% CRP could support between £45bn and £90bn of new business, and therefore investment from the insurance sector. There is inherent uncertainty around these estimates, not least because firms could choose to return some of the released capital to shareholders.^{footnote [12]}

48. The wider reform package enabled by putting the MA on a sound footing would facilitate investment in long-term productive assets. Taken together with the proposed changes to MA eligibility rules and supervisory processes discussed in HMT’s consultation, the package of reforms would make it easier for insurers that wish to place these new assets or redeploy their existing assets into investments which support growth, infrastructure and the transition to net-zero. The PRA has taken these factors into account in its analysis, as they relate to the matters to which it must have regard when making policy.

49. Combinations of reforms which reduce overall capital levels from their starting position will inevitably lead to some reduction in financial resilience, and this is directly relevant to the PRA’s statutory objectives for safety and soundness and policyholder protection. If liability values are below market transfer values, the risk of an insurer failure being disorderly would also increase. Assessing the quantum of these effects is inherently challenging.

50. In the round, it is the judgement of the PRA that the decrease in financial resilience which would result from a package consisting of a c.60% risk margin reduction and an FS with a CRP equivalent to at least 35% of credit spreads would still deliver an outcome in line with the PRA’s statutory objectives, taking into account the level of capital released, available market data on transfer values, and noting the potential positive benefits which could accrue from the reforms. A package of this nature would be within the potential ranges noted in HMT’s consultation. However, there are other combinations of reforms that would result in outcomes that the PRA considers would not be compatible with its statutory objectives.

Annex

• Solvency II Review: Matching Adjustment and reforms to the fundamental spread