Our response to the HMT Call for Evidence is available here. The executive summary is reproduced below for ease of reference.
A more transparent and principles-based regime
In the current Solvency II regime, there is a substantial degree of disconnection between what things mean and how they are calculated. For example, as the Call for Evidence notes, it is generally accepted that the method by which the Risk Margin (RM) is calculated does not lead to a sensible outcome in relation to its intended meaning, both in relation to the size of the RM and its volatility in response to changes in interest rates. Similar effects apply in relation to the Matching Adjustment as discussed later in this note. The Call for Evidence rightly identifies an opportunity to rebalance the mix of rules and judgement.
This is likely to result in the PRA having more powers, both formally (through changes to legislation or the PRA Rulebook), and informally through a greater degree of establishing and interpreting principles and exercising judgement. This increase in powers should be balanced by an increase in transparency, so that the firms the PRA regulates and supervises understand the reasoning behind the PRA’s expectations and where appropriate are able to challenge the PRA’s judgements effectively.
As part of improving transparency, key aspects of the regime should be better defined, for example by explaining what an element is intended to represent in terms of a risk appetite or confidence level. The SCR is well-defined in these terms (99.5% confidence over 1 year), but the same cannot be said of the Fundamental Spread, where in the current regime the level of confidence relating to long-term default and downgrade risk is an opaque consequence of the methodology used to calculate it. This opacity is unhelpful, for example it makes it difficult to ensure that different asset classes are being treated in an equivalent way.
Risk Margin
We agree with the diagnosis in the Call for Evidence of the issues with the RM. The issues are sufficiently serious that technocratic adjustment to the calculation methodology (for example the ‘lambda factors’ recently proposed by EIOPA[1]) is insufficient and a more fundamental reform is needed.
The MOCE concept introduced by IAIS[2] appears to be more sensibly designed than the RM and is based on well-defined confidence levels for life and general insurers. Aligning with IAIS would have obvious practical benefits for firms that will become subject to IAIS in due course.
Matching Adjustment
We support the economic rationale for allowing a mechanism like the MA. As the Call for Evidence notes, there are many benefits from the MA. However, there are pros and cons to any such regime. We consider the MA regime needs to be revised to improve outcomes for policyholders, firms, the PRA, and the UK economy.
The MA should be regarded by firms as a privilege, not a right. Both firms and the PRA should approach the MA with an appropriate degree of humility, as it relies on accurate risk measurement over long time horizons.
We consider the definition of the MA as an adjustment to liability valuation to be unhelpful. In our view, it is better to frame the Matching Adjustment as a way of recognising the risk-mitigating effects of a credible hold-to-maturity strategy. It should be shown as a separate item on the asset side of the balance sheet, arising from a synergy between the assets and liabilities and depending on both, but not directly part of the valuation of either. In this framing, the investment assets and insurance liabilities both remain at market / market-consistent values. The separate MA asset would recognise the risk-mitigating effect of the hold-to-maturity strategy on the regulatory balance sheet, net of all relevant risks, including but not limited to those measured by the current FS. This framing more clearly expresses what the concept means, rather than how it is calculated (via an addition to the liability discount rate).
The MA relies on separation of credit from non-credit risk. This is technically challenging, and impossible to achieve with complete confidence. There is a risk of falling into a ‘best is the enemy of the good’ trap here: lower annuity prices, more long-term investment and reduced procyclicality are all very worthwhile outcomes, and our view is that these outcomes merit accepting some degree of uncertainty over the measurement of credit risk. The balance to be struck is a risk appetite question for HMT and the PRA.
The credit-riskiness of the cashflows is not the only risk. All risks to the credibility of a hold-to-maturity strategy need to be considered, either through measurement (in the equivalent of the FS) or exclusion (the eligibility criteria). The presentation of the MA as an asset rather than an addition to the discount rate enables the equivalent of the FS to be composed of different monetary reserves. This is a more general lens through which to examine the allowance for risks and facilitates a more flexible approach than either excluding risks entirely or forcing them through the FS as a basis points item. We believe it would then be possible (within reason) to relax eligibility criteria around cashflow fixity, prepayment risk and cashflow matching by introducing compensation for these risks within an expanded definition of the FS.
The current FS is based on a complex, inelegant, and internally inconsistent methodology inherited by the PRA from EIOPA. In our view, the definition of the credit element of the FS (as distinct from prepayment and mismatching elements) should be recast as:
- a best-estimate allowance for all relevant risks impacting on asset cashflows; plus
- an additional well-defined allowance for risk, which could be considered as a ‘credit risk premium’.
The allowance for the credit risk premium should be based on a well-defined risk appetite articulated by HMT or the PRA, with a transparent and detailed methodology giving effect to this risk appetite published by the PRA so that interested parties can scrutinise and challenge the calibration against this appetite. This risk appetite would in effect express the balance struck between the trade-offs embedded in the use of the ALMA concept, reflecting costs, benefits and risks to policyholders and wider society.
The overall calibration of the expanded FS is again a risk appetite question for HMT and the PRA, within the design considerations outlined above. The balance between using credit ratings and market prices needs to be reconsidered: the MA regime essentially gives almost no weight to current market prices as a source of information on credit risk, other than in the special case of the BBB cliff in Article 77c(1)(c) of the Solvency II Framework Directive. The focus on credit ratings has a very strong stabilising effect, arguably excessively so: in our view this needs to be rebalanced towards giving some weight to market prices alongside credit ratings. Again, this is primarily a risk appetite question, as it bears on the cost/benefit trade-off from the MA.
In the current MA regime, we believe there is more strength in the SCR than on the base balance sheet. This balance should be reconsidered: we believe it would be better to have more strength in base than the current FS, particularly for internally rated assets, and a corresponding reduction in the SCR – when the base is stronger, there is less risk to be covered by the SCR. While there would be a loss of diversification benefit from reducing the SCR, there would also be a reduction in the capital buffers firms hold on top of the SCR and the impact might reasonably be broadly neutral overall.
Solvency Capital Requirement
The distinction between the Standard Formula and (partial) Internal Models is too hard-edged and more flexibility is needed. This should include:
- More granularity in the risk modules – for example at present credit risk is ‘all or nothing’ and it would be beneficial to some firms to be able to use the SF for corporate bonds while having an internal model for risks where the SF was inappropriate. This would be a more proportionate approach for small firms with specialised exposures.
- Widening the scope of undertaking-specific parameters (USPs), in particular to include asset risks. This would enable customisation within a predefined methodology for asset classes other than corporate bonds, and act as a helpful half-way house between the SF and a full IM.
There is limited transparency in the calibration of the SF at present. The PRA should improve this on a gradual basis starting with the most material risks. This would also have the benefit of improving the transparency of the PRA’s Quantitative Indicator (QI) framework[3]: if the historical circumstances had been such that the PRA had adopted its own view for SF purposes there would be no need to have distinct IM QIs for the same risks, although some QIs would still need to be adapted to fit the risk profile of each firm applying for an IM.
The PRA should have powers to adjust Technical Provisions directly (in the manner of a ‘capital resources add-on’ under the prior UK ICAS regime) rather than using an adjustment to the SCR to achieve the same effect. The PRA should be accountable for the use of this power via existing mechanisms, including being subject to the Tribunal, and should also disclose aggregate statistics on how often it is used.
[1] See page 30 of EIOPA’s Opinion on the 2020 Review of Solvency II at https://www.eiopa.europa.eu/content/opinion-2020-review-of-solvency-ii_en
[2] See section 5.3 of Public 2020 ICS Data Collection Technical Specifications at https://www.iaisweb.org/page/supervisory-material/insurance-capital-standard/file/90757/public-2020-ics-data-collection-technical-specifications
[3] See the 2016 letter from Sam Woods (then in his capacity of Executive Director of PRA Insurance Supervision) at https://www.bankofengland.co.uk/-/media/boe/files/prudential-regulation/letter/2016/sam-woods-reflections-2015-solvency-ii-internal-model-approval-process-jan-2016.pdf