SPOTLIGHT ON RISK TRANSFER AND INSURANCE SERIES | PART 4
Market risk capital
requirements of superfunds
and insurance buyout firms
Spotlight on risk transfer and insurance series | Part 5
Recent updates from both the Department of Work and Pensions (DWP) and The Pension Regulator (TPR) have confirmed that superfunds continue to play a significant role in government plans for the DB pension risk transfer market, while also clarifying DWP’s intentions for a specific quantitative standard for the level of regulatory capital that superfunds should hold, thereby establishing an intended regulatory capital differential with buyout firms. Here we discuss the intended regulatory capital differential, and how further evolution of the regulatory capital approach for superfunds may be necessary to achieve DWP’s objectives.
Superfund security and pricing in the DWP Consultation Outcome
The DWP Consultation Outcome reiterates the Government’s previously stated position that it does not intend for superfunds to provide an additional option to DB pension schemes that can currently afford, or almost afford, an insurance buyout. Superfunds are to provide a cheaper option for those schemes that are unlikely to reach a buyout funding level in the foreseeable future. The CO envisages a superfund entry price of approximately 90% of the price of an insurance buyout.
The CO also recognises that superfunds’ lower entry price must be accompanied by a lower level of benefit security relative to buyout: superfunds cannot be expected to capitalise in line with Solvency II (or its forthcoming Solvency UK version) whilst at the same time offering a materially more affordable price than buyout.
The CO specifies the security level that it intends for superfunds: a 2% probability of failing to pay all benefits is an acceptable level of superfund security for DWP.
However, the CO does not propose that the superfund capital requirement should be set with direct reference to this 2% probability of funding all benefits in a run-off projection. Instead, it proposes to deliver this level of security through a capital standard that is similar in form to Solvency II, but slightly weaker in parameterisation: whilst Solvency II requires insurance firms to be capitalised to a 99.5% one-year Value-at-Risk (VaR) standard (i.e. a one-in-200 year event), the CO sets out that superfunds should be capitalised to a 99% one-year VaR standard (i.e. a one-in-100 year event).
Moving from Solvency II’s 99.5% one-year VaR requirement to a 99% VaR requirement could reduce solvency capital requirements (in excess of technical provisions) by a proportion of 10% to 15%1 (i.e. if technical provisions were 100 and the Solvency II Solvency Capital Requirement (SCR) was 10, this change in capital requirement standard could be expected to reduce SCR from 10 to around 8.5 to 9).
This reduction in capital requirements, on its own, is unlikely to induce a 10% reduction in pricing – a back-of-the-envelope cost-of-capital modelling calculation implies the buyout price impact of this reduction in capital requirement would be more like 1%. This suggests DWP’s aspiration for superfunds to price at 90% of a buyout price will not be induced solely by its intended regulatory capital differential between superfunds and BPA firms. However, superfunds may also anticipate cost efficiencies driven by differences in operating and administration models relative to buyout firms, and these may drive further price differentiation. Superfunds will also face less demanding re-capitalisation requirements than those that apply to buyout firms. In practice, the differences in buyout and superfund pricing may be more nuanced. For example, the pricing differential may tend to be greater for deferred members than for pensioners.
TPR’s updated guidance confirms that, for the interim period until DWP’s new legislative program for superfunds is put in place, the TPR expects the market risk element of a superfund’s risk profile to be supported by a capital buffer that will ensure there is a 99% probability that total assets will exceed a minimum level of technical provisions after five years (so, a five-year 99% VaR instead of the one-year 99% VaR planned by DWP).
The revised guidance updates the discount rate for this minimum level of technical provisions from gilts + 50 bps to gilts + 75 bps. Whilst this represents a weakening in the TPR’s minimum technical provisions basis, the new discount rate is unlikely to be materially weaker than the basis implied for insurance technical provisions under the Solvency II Matching Adjustment (MA) requirements that usually applies to buyout firms’ liabilities.
We noted above that a shift from 99.5% to 99% one-year VaR would likely reduce capital requirements (in excess of technical provisions) proportionally by some 10%-15%
The relationship between a 99% one-year VaR and a 99% five-year VaR is less straightforward – the difference between these two capital requirements will depend on the relative levels of the expected return and risk that are assumed for the assets (the higher the expected return, the more likely it is that moving from one-year to five-year VaR will reduce rather than increase the capital requirement) and the assumed ‘time structure’ of the risk profile (for example, the assumption of material mean-reversion in returns will tend to have the effect of reducing measures of longer-term risk relative to shorter-term risk horizons).
The recent modelling analysis produced for TPR by Mercer showed that, for a gilts + 0.75% technical provisions basis, the shift from 99% one-year to 99% five-year 99% VaR could produce higher or lower capital requirements, depending on the assumed asset strategy (this result will also tend to vary over time as most models will assume that the expected return of credit-risky assets varies with the prevailing level of credit spreads). For the asset strategy that was intended to represent a typical insurance buyout firm’s portfolio, the capital buffer produced by the 99% five-year VaR was very slightly less than what would be required by the 99% one-year VaR.
This result, together with the observation that the 99% one-year VaR will be less than the 99.5% one-year VaR, suggests that the 99% five-year VaR capital requirement for the indicative buyout portfolio would, all else being equal, be some 10% to 20% lower than the one-year 99.5% VaR capital requirement.
However, there is an important aspect of the TPR basis that could have the effect of strengthening the superfund capital requirement relative to the SII MA standard. When calculating an SCR for Matching Adjustment business, Solvency II permits the technical provisions discount rate to be materially increased when credit spreads are stressed upwards.
Adjustments to the discount rates
The rationale for this upward adjustment in the liability discount rate is that at least some of the credit spread increase will usually merely signal an increase in expected asset returns rather than higher expected credit losses. As long-term buy-and-hold investors with high quality asset-liability cashflow matching, buyout firms’ MA asset portfolios should therefore expect to recover this element of the spread stress-induced market value losses over the remaining term of the assets. Allowance for this effect provides a partially offsetting reduction in liability valuations in stress, which has the effect of materially reducing the market risk SCR for MA business.
The form of this adjustment is prescribed in the Standard Formula. All UK buyout firms currently use an internal model for this element of their SCR and each internal model will have its own methodology and calibration that will produce results specific to the given firm. However, it is standard industry practice for these models to generate a similar effect to the Standard Formula by incorporating a dynamic adjustment to the MA liability discount rate that is assumed in the stresses that determine the SCR.
Although data analysis has already completely revolutionised our everyday retail journeys, in the workplace many key decisions are still based purely on anecdotal evidence or instinct alone. Slowly but surely, the juggernaut is turning and analytics is on the rise in the workplace. Indeed, we are heading towards a new destination where Employer DNA will deliver sustainable, robust and innovative strategies.
"We are heading towards a new destination where Employer DNA will deliver sustainable, robust and innovative strategies."
I started by saying that “DNA is the very material that defines our uniqueness – the very substance that carries the information we need to survive and to thrive.” The same is true for Employer DNA. As you work your way up the rungs of the data analytics ladder, the closer you get to the top, the more you’ll realise that your Employer DNA really does contains the insights you need to both survive and to thrive.
This effect provides a material offset to the market risk capital requirement under MA. For example, consider a five-year A-rated zero-coupon corporate bond that offers a spread over the risk-free yield of 100 basis points. In the MA, this asset would attract a Fundamental Spread of around 28 basis points, implying an MA discount rate of risk-free + 72 basis points. In the Standard Formula, five-year A-rated corporate bonds incur a capital charge of 7%. This is equivalent to a credit spread stress of 155 basis points for the five-year bond. For MA business, the Standard Formula allows the MA liability discount rate generated by this asset to increase by 62 basis points in this stress, resulting in a liability discount rate of risk-free + 134bps being used in the stress. This increase in the liability discount rate used in stress reduces the credit SCR from 7% to around 4.1% (proportionally, a c. 40% reduction in SCR).
The TPR requirement for the technical provisions discount rate to be at least as strong as gilts + 0.75% appears to rule out this dynamic liability discount rate offset feature that is a core element of the Solvency II Matching Adjustment SCR. This could result in superfunds’ market risk capital buffers exceeding the market risk SCRs of buyout firms.
This may help explain why the market risk capital buffers produced in the Mercer modelling analysis tend to be a bit bigger (rather than smaller) than the market risk SCR levels typically associated with buyout business. For example, on the gilts + 0.75% basis, the modelling analysis shows that the buyout-style investment strategy produced a capital buffer of almost 12% of technical provisions. The market risk SCRs of buyout business would typically be materially lower than this and of the order of 8%-10%2 (there may be other factors that also contribute to these differences, such as Mercer’s specific asset modelling and investment strategy assumptions, but it is likely that the absence of a liability discount rate ‘offset’ makes a material contribution to the assessed superfund capital buffer3.)
Capital modelling governance
Both Solvency II and the TPR guidance permit a principle-based
approach to the assessment of market risk capital requirements.
However, the two governance frameworks that regulate the use
of capital models in the assessment of risk-based capital
requirements are likely to have some notable differences.
The TPR approach is set out in Appendix B of their interim guidance. In summary, superfunds will be required to disclose the details of their capital modelling to the TPR, including justification for assumptions and how these have been validated and authorised, and the results of sensitivity tests of key assumptions. TPR will review the modelling and may wish to investigate specific aspects of the modelling further. TPR expects to be informed in advance of any material changes to the capital model and the reasons for those changes.
The path to principle-based capital assessment for insurance firms is more formalised than the interim regime outlined by TPR. This is partly a result of Solvency II’s established legislative framework and the presence of a Standard Formula alternative to a firm’s internal capital assessment. Under Solvency II rules, a firm must apply for permission to use an internal model to calculate the SCR, and it must be formally approved by the relevant national supervisor (PRA in UK) before it can replace the Standard Formula as the firm’s SCR calculation method.
Insurance firms are now well-accustomed to the technically rigorous review processes that PRA undertake prior to approving an internal model for use in SCR assessment. Meeting the PRA’s expectations for the credit risk modelling of MA portfolios is a particularly technically demanding task. This is partly an inevitable result of the innate complexity of Solvency II’s MA framework. But it also reflects the importance that the PRA attaches to highly rigorous capital assessment methods. A substantial amount of technically skilled actuarial resource is deployed by the PRA in robustly reviewing the proposed internal models of insurance firms and ensuring they are of a high technical standard.
It is not yet clear what level of technical scrutiny TPR intends for its model reviews: for example, will it seek to build a technical capability similar to that of the PRA in order to perform its reviews of superfund capital models? Or will it publish guidance with detailed modelling expectations that are analogous to PRA supervisory statements?
There is also the question of consistency of modelling strength between superfunds and buyout firms.
DWP has used a Solvency II-style capital standard for superfunds to specify a security level for superfunds that is easy to compare with that of buyout firms. But this intended relative strength will only be achieved if there is consistency in capital modelling standards between superfunds and insurance buyout firms. At present it is unclear what mechanisms, if any, will be used to achieve this cross-sector consistency. The above discussion of the likely capital impact of the lack of a liability discount rate offset in stress for superfunds highlights that a lack of consistency in technical modelling standards may have inadvertent and material impacts on the relative strength of the capital standards of superfunds and buyout firms.
In summary
The relative capital treatment of financial market risk within superfunds and buyout firms will be a critical driver of the price and security differentiation that DWP is seeking.
It is also an area with the potential for considerable technical complexity. The new superfund regulatory regime will require careful implementation to deliver the degree of regulatory capital differentiation between superfunds and buyout firms that is sought by DWP. This blog has explored some of the key issues that may arise in the implementation of the intended regulatory treatment of financial market risk in superfunds. In particular, the discussion noted that:
- The difference between DWP’s stated superfund capital standard and the Solvency II capital standard is unlikely to, on its own, induce a 10% reduction in pricing relative to buyout.
- The TPR requirement for the technical provisions discount rate to be at least as strong as gilts + 0.75% appears to rule out the dynamic liability discount rate offset feature that is a core element of the Solvency II Matching Adjustment SCR. This could result in superfunds’ market risk capital buffers exceeding the market risk SCRs of insurance firms.
- DWP has used a Solvency II-style capital standard for superfunds to define a capital requirement for superfunds that is easy to compare with that of buyout firms (and that is demonstrably lower). This intended level of relative capital strength will only be achieved if there is consistency in capital modelling standards between superfunds and insurance buyout firms. A lack of consistency in technical modelling standards may have inadvertent and material impacts on the relative strength of the capital requirement levels of superfunds and buyout firms.
References
1 - The ratio of the 99.5th to 99th percentiles of the standard normal distribution is 90.3% (2.326 / 2.576). This suggests the first-order impact of the change in capital standard would be to reduce the SCR by around 10%. The use of a fatter-tailed distribution such as, say, a t distribution with 6 degrees of freedom would imply a reduction in SCR of 15%.
2 - Market risk SCRs are disclosed in firms’ Solvency and Financial Condition Reports (SFCRs), although a breakdown between buyout and non-buyout business is not usually provided. At end-2022, buyout-focused firms such as PIC, Rothesay and Just reported market risk SCRs of 10%, 7% and 8% of their technical provisions respectively.
3 - We should note that the scale of the liability discount rate offset effect will reduce as the length of the VaR projection horizon is extended (as the end-period technical provisions that are being valued with the discount rate will reduce as projection horizon is extended). However, it will still be highly relevant to a 5-year VaR as pension liabilities’ duration will tend to be materially longer than 5 years. Furthermore, DWP intends to ultimately use a 1-year VaR projection horizon.
