137.Both written and oral evidence to the previous Treasury Committee highlighted the lack of proportionality both within the PRA’s application of Solvency II, but also the rules-based approach of the underlying legislation. For example, regarding the genesis of Solvency II, the Institute and Faculty of Actuaries stated that:
“The early simplicity of concept was distorted in the process of negotiation, both of the original Directive, the amendments under the ‘Omnibus II’ Directive and in the Commission’s Delegated Regulation.
‘’Solvency II has encouraged a rules-based and bureaucratic implementation, partly because of the Lamfalussy process followed: with the Directive set in stone early on, and Level 2 text (providing detail) following much later. This has meant, at times, that decisions have been made in terms of interpreting the Directive text as a rigid truth rather than on the merits of the case.”
138.The ABI supported this evidence, noting the volume of regulation that firms are expected to follow:
“[I]n practice the regime and its requirements are overly complex and detailed. This needs to be addressed to allow for a more proportionate, effective and sustainable prudential regulatory regime. The volume of text alone is indicative of this: approximately 300 pages of level 1 text, approximately 2000 pages of level 2 text, with approximately 975 more pages of EIOPA guidelines—totalling over 3,200 pages.”
139.Although in theory the legislation allows for ‘proportionality’, in practice it encourages a detailed rule based approach to implementation which a cautious and professional regulator such as the PRA found it difficult to avoid. KPMG compared the UK approach with other European Union countries:
‘’Based on our discussions with KPMG colleagues in other EU countries, and with pan-European clients, the implementation of Solvency II in the UK was a more difficult and onerous process than in other EU countries. In particular, the approach to obtaining Internal Model approval was more onerous than we understand was the case in other EU countries.’’
140.Two specific examples of proportionality highlighted by the majority of submissions are in relation to capital models and data requirements.
141.The rules-based approach taken by the Solvency II legislation has led to an excessively rigid approach to capital models, particularly Internal Models. Capital requirements under Solvency II can be calculated using the Standard Model or, subject to regulatory approval, an Internal Model. Whereas most firms by number use a Standard Model, almost all the larger or more sophisticated insurers use (or will have a preference for developing) their own Internal Model, which can be more complex and tailored to the circumstances and risks of the individual firm.
142.The use of Internal Models was encouraged by the PRA. To prevent misuse, an Internal Model’s structure and assumptions have to be agreed by the regulator before use. The nature and sophistication of the UK insurance industry, combined with the PRA’s approach, meant that the PRA had to approve more Internal Models than any other national regulator, and necessarily had to develop a structured process. The development and approval of Internal Models in the UK has come under significant criticism from many respondents because of the work and cost that approval it entails, but also because of the resulting lack of flexibility. The Institute of Actuaries summarised the weaknesses of taking an approach that builds everything around one central, but complex, model often with thousands of assumptions:
‘’The development of Solvency II has been more model dependent than it should have been. Belief in the benefits of capital models reached a high water mark in 2007–8 when Solvency II was being cast and although there have been more considered views recently, development has encouraged an excessively rigorous testing of models. There is then an expectation of complete reliance on such models, rather than less exacting testing of models and more flexibility in their use, which may be a better approach.”
143.Even the most sophisticated or complex models can only ever provide an approximation of reality. Models are also far more likely to be used if they produce sensible results in which the insurer has faith. Different models have different uses and may be appropriate for varying purposes. For example, a far more granular model would be necessary for looking at one aspect of the business, compared to a model assessing the overall firm. The current approach fails to recognise this because it emphasises the importance of using the (single) capital model for a whole range of purposes.
144.The PRA requires approval for the use of Internal Models to calculate capital requirements, and further approval of any subsequent amendments to those models. The focus on a large complex Internal Model has been accompanied by a detailed and expensive validation process. The sentiments of the IFoA submission were echoed by many other submissions to the previous Committee:
‘’In our view the process developed by the PRA to review and approve Internal Models is too onerous and time-consuming, and in turn, this has led to excessive levels of Internal Model validation within firms. Whilst some validation and review is clearly necessary, we believe the level of validation and documentation required has been excessive and, at the margin, of limited value. This approach does not seem to be sustainable and is driving unnecessary cost into insurers, which may ultimately be to the detriment of consumers.”
The IFoA noted the effect that this process has had on the regulator:
“Internal Models have also left the regulator very careful and cautious about what it is agreeing to. The PRA is having to sign-off on an ongoing model as opposed to a ‘point in time’ capital figure, which then results in them being—understandably—risk averse’’.
145.Of particular concern for those insurers whose models are now approved is that the process for gaining approval of changes to their models is also onerous and is expected to take several months, constraining their commercial flexibility, and meaning that they will be required to use either inappropriate or unapproved models in the interim.
146.Finally, some witnesses argued that if the Standard Model were refined, it would be applicable to more insurers, reducing the need for complex, and costly, Internal Models. When questioned by the previous Treasury Committee on this point, Andrew Chamberlain of the IFoA suggested that it “would certainly be a laudable thing to achieve”, adding “ a number of rigidities” make it “impossible for many” to operate, particularly in the UK. These views were echoed by Jane Portas of PwC, who noted there were “opportunities” to refine the approach. This may be of limited significance to currently authorised insurers who have made their decisions and are unlikely to change in the short term, but it does have an impact on the attractiveness of the UK regime for new entrants.
147.Written evidence to the previous Committee presented suggestions for changes to the standard model. For example, Steve Dixon of Steven Dixon Associates LLP advocated:
“More research on the standard formula on two key aspects: the operational risk module and the morbidity risk module for health insurance. The former arrives at amounts of capital requirements that appears too low for the majority of smaller firms and too high for large firms. The latter generates a capital requirement to cover 2 or 3 times the expected sickness by the additive effect of the combination of a stress on incidence of periods of sickness and a stress on the speed at which people recover from sickness.”
While KPMG asserted that:
“The Standard Formula [Solvency Capital Requirement] does not cope well with the following risks, often due to a simple lack of clarity: Pension scheme risk…Operational risk…Sovereign debt risk…Currency risk…Deferred tax…The Standard Formula has no volatility stresses in it. This is in our view an omission, particularly in respect of equity volatility which is material for many life insurers.”
148.In their supplementary evidence following their oral evidence session, the PRA identified the Internal Model change process as an area where they are open to reform, and will explore options with the ABI:
“The PRA will continue to review the ongoing appropriateness of model approval processes and ensure they do not impose an excessive burden, whilst respecting the constraints of the Solvency II regime. We are currently reviewing a large number of Internal Model and other approvals and acknowledge the resource-intensive nature of the regime. We are actively looking at ways to reduce the burden on supervisors, firms and boards. It should be noted that although, under the relevant Solvency II provisions, the PRA has six months from the date of application to provide a decision on major model changes, our experience so far has shown we have taken considerably less than six months to provide decisions.”
149.The Committee notes the widespread and consistent concerns expressed by firms over the proportionality of the PRA’s approach, particularly with regard to the review and approval of internal Models and amendments to those models. The PRA is urgently asked to review its practices and report to the Committee on proposed changes.
150.The current EU legislation allows for proportionality in national regulators’ approach to capital models. There should be no reason to wait until the UK has left the EU to take advantage of these provisions. The Committee requires the PRA to note this stipulation and report on intended actions.
151.Evidence suggests that the Standard Formula does not adequately reflect the risk profile of many firms, so they will need to develop their own internal models to run their business, even if those models that are not PRA-approved. Examples of areas where the Standard Formula is insufficiently sophisticated include interest rate risk and longevity risk, particularly where a firm has products which contain financial guarantees which require a more sophisticated evaluation. Firms can also be exposed to market risks—such as inflation risk and gilt spread risk—where liabilities are hedged using gilts but the discount rate which must be used for valuing liabilities must be swap based. These are not covered by the Standard Formula. Its allowance for operational risk is formulaic and does also not reflect the true risk profile of any individual firm. It is likely to understate the risk for a sophisticated firm, although the impact of this can be offset by the fact that it does not allow for the benefits of diversification.
152.The PRA should also assess whether the Standard Formula could be enhanced for the benefit of existing users, especially if an improved version could save some firms from needing, or choosing to upgrade to, an Internal Model. The Committee suggests that the PRA consider the specific ideas for improvement put forward by witnesses in evidence to this inquiry.
153.Regular reporting is of a different order of magnitude from that under the previous regulatory regime. Annual and quarterly returns requiring immense detail are obligatory. To these the PRA has added UK requirements which the ABI said were “over and above the EU requirements”, which many respondents claimed had imposed a “very significant UK-specific burden”.
154.It is unclear what value much of the reporting adds. KPMG stated that some of the Quantitative Reporting Templates “are in our view of limited benefit”, while Legal & General states that “the disclosure requirements…are excessive and of limited use to regulators, investors, intermediaries and policyholders.
155.The industry is concerned with the “significant ongoing cost”, but there is also the danger that the “sheer volume of the rules and guidance acts to limit understanding and effectiveness of Solvency II”. Lloyd’s of London make the following remarks in their submission:
“Solvency II’s requirements are very detailed and it is not clear what use supervisors can make of all the information they receive. The reporting should enable supervisors to get maximum benefit which would be better achieved by reducing the volume and frequency of template collection, which would also reduce the administrative burden on insurers.”
156.One consequence of the additional data requirements is an impact on competition. For example, the ABI noted the disproportionate impact on small firms, because, although they may “apply for an exemption to some, but not all, quarterly reporting requirements…the PRA has also set out an expectation that such firms should be able to respond to ad hoc requests from the PRA for any of these exempt templates”. This sentiment also has traction outside the UK, with the Insurance Europe’s Reinsurance Advisory Board concluding that a “‘one size fits all’ approach is unworkable, as it results in an approach whose complexity is inappropriate for companies with smaller and simpler risks and leads to results that are misleading or wrong for undertakings with larger and more complex ones”.
157.Following their oral evidence the ABI, Legal & General and Lloyd’s of London submitted further evidence to the previous Committee on the level of burden. The ABI stated that in comparison to the previous regime volume had increased by 4–8 times, frequency had increased from annually to quarterly and deadlines are reducing over the 3-year implementation period. Meanwhile Lloyd’s of London explained that it must collect information from each of its 111 syndicates on a quarterly and annual basis, meaning that the “reporting burden is huge”. For this reason Lloyd’s of London advocates several changes, including discontinuing the quarterly reporting in favour of half yearly reporting, which would be consistent with the frequently of insurers’ financial statements. The ABI has recommended four specific changes in this area.
158.The regulator made several arguments concerning the volume of regulatory reporting. In oral evidence to the previous Treasury Committee, Sam Woods argued that the data “is useful” as during times of a stress supervisors can tell him “very rapidly” whether or not a firm is exposed to a particular risk. He said “that is really going to matter when, hopefully far in the future, we come to bail in a bank”. Victoria Saporta added that, while firms should know their counterparties and be able to produce this data as soon as possible, in her experience “that is not always the case”, notably during the recent crisis. Sam Woods later noted that “the law is pretty clear that we just have to collect some of this stuff,” although the then Chairman of the Committee retorted this was “up to a point, Lord Copper”.
159.However, the regulator did acknowledge that this issue was “probably the most consistent theme across the shop”, noting that quarterly reporting does “feel rather onerous” given the limited turnover of assets, while total costs were estimated at between £200 million and £250 million industry-wide.
160.For these reasons the regulator has already taken some action, notably a waiver for small firms which “takes them out of about 70% of the quarterly reporting burden”. They also stated that they are keeping an “open mind” on the current requirements. Thus, while the numbers inform the “basics of supervision”, and indeed they cannot “drop [requirements] entirely”, Sam Woods and David Belsham acknowledged that they should undertake a review to “see whether [further] waivers can be given” and whether some of the data could be collected by exception. He subsequently wrote to the Committee to confirm that he would review the PRA’s implementation of the Solvency II reporting requirements, with specific reference to the usefulness, or otherwise, to the PRA of the data collected, and the impact on firms of the reporting requirements.
161.There is evidence that the regular reporting requirements are costly and overly detailed. The Committee believes that they could be streamlined, reducing the burden and cost on firms, and reducing the risk that the PRA could miss something. The PRA should review the information collected—both in the EIOPA and national templates—and assess it in the context of both usefulness and cost effectiveness, notably the quarterly data. The merits of exception reporting and/or focusing more on changes to previous information should also be considered.
161 Two thirds of respondents to Grant Thornton’s market survey (October 2016) thought that Solvency II is too complicated and two thirds also agreed that the costs of Solvency II were disproportionate to the size of the business
174 For example: SOL026 paragraph 83; SOL008 paragraph 4d; SOL006 paragraph 4d.
188 Q139 and SOL050
202 Qq203, 205
25 October 2017