## Insurance supervision at the PRA
I am grateful to Dan Curtis for his help in preparing this speech.
1
Speech given by Prudential Regulation Authority
Sam Woods, Deputy Governor, Prudential Regulation and Chief Executive Officer,
London Business School 20 March 2017
Good morning.
The UK has a world-leading insurance sector which provides a wide range of vital services to the real economy and society. It promotes growth and employment, has proved its mettle in turbulent times and the country can be proud of it.
A necessary precondition to such success is world-class insurance supervision. That is what we aim to deliver at the PRA. Like insurers, our job is to be forward-looking and prudent about risk. Where we differ from private firms is that we do this motivated solely by the public interest.
The recent debate about insurance regulation - while timely and important - must seem to some like a cacophony of acronyms, statistics, models and assumptions. But strip this back and you'll see there is an essential, irreducible human core to it all. Some of the oldest and most vulnerable in our society have invested their life savings into long-term annuity contracts. By pooling and transferring many kinds of risks - from cyber to marine - insurers provide cover which is essential for economic activity. And by protecting for critical illness or personal accident, insurers commit to being there when you need them most.
So when we talk about promoting insurers' safety and soundness, and protecting their policyholders, this is what we have in mind.
The PRA has squeezed much into its short life since it began supervising insurance companies in 2013, not least the enormous undertaking of implementing Solvency II, and knitting insurance supervision into the wider Bank, where there are obvious connections with areas like monetary analysis.
And today, the Bank of England's Independent Evaluation Office has published its thinking on the PRA objective which is unique to insurance, of contributing to the protection of policyholders.
All this makes now the ideal time to refresh our approach to insurance supervision.
I will begin by taking stock of insurance regulation . These are the rules and standards which apply across the board, and have been recently the subject of ongoing and lively debate. Now that Solvency II has applied, how do we take it forward?
From there I would like to explore our approach to insurance supervision - the day-to-day business of understanding, assessing and challenging the 469 1 firms we are responsible for. In particular, how do we at the PRA think about policyholder protection?
1 This is a consolidated view of insurance groups - there are approximately 600 insurance entities, including solos that are part of groups.
2
And what is the point of all this? Our aim is to maintain a resilient insurance sector which does not pass risks back to policyholders when they crystallise.
## Solvency II: a stocktake
For the uninitiated, Solvency II is a new regulatory regime for insurance companies across the EU, introduced on 1 January 2016 after years of wrangling. There is now a lively debate about how the new regime is working out here in the UK. It seems that almost every day the specialist press contains new opinion pieces from industry insiders or consultants about how Solvency II is constraining their business or turning out not to be as bad as feared. There is also the important review by the Treasury Select Committee, at which I gave evidence recently. And very occasionally others outside the magic circle of insurance enthusiasts (of which I count myself one) take a passing interest. Inevitably, on a subject as complex as Solvency II, there will be a wide range of opinions. And I think that a degree of constructive tension between regulator and regulated firms is a sign of a healthy and properly functioning relationship. But in any debate it pays to consider both sides.
So what does the insurance industry think about Solvency II? Unsurprisingly, there is a range of views. Lord Turnbull, former board member at Prudential, was unequivocal in his comment that Solvency II was 'an absolutely dreadful piece of legislation'. Others, particularly those working in general insurance, are more sanguine - for instance John Parry of Lloyd's of London told the TSC that 'in terms of Solvency II…damaging competitiveness or returns on capital, no, we have not seen that'.
Written submissions from the industry to the TSC generally reveal two themes. First, having just consumed this sizeable piece of new regulation, the industry has no desire whatsoever to sit back down at the regulatory table and consume another, radically different, meal. By my reckoning, only one submission to the TSC calls for a fundamental change of regime. And second, that while the PRA has done a reasonable job in steering everyone across the implementation line, it has in the view of some parts of the industry been over-zealous in places. I'll comment on each of these themes in turn.
## Solvency II indigestion
I very much agree that having another big meal at this stage would be most unwise. However, I acknowledge that the size and variety of the Solvency II meal has left all of us facing some degree of indigestion. When faced with live market conditions and commercial realities, some aspects of the regime have been found wanting. So I agree with most industry commentators that it is in nobody's best interests - not firms, policyholders or regulators - to make wholesale changes. But that does not mean that we cannot make a series of tweaks and improvements to the overall system, by way of a digestif.
3
For example, there has been considerable criticism of the Risk Margin - a new feature of Solvency II which aims to allow for uncertainty in liability valuations - and I agree that its current design and implementation does not achieve the intended goals. The concept of a risk margin is a sensible one, but its current implementation in Solvency II is flawed. Because the root causes of the problems are baked into European legislation, we need to fix them at source, and are actively and constructively engaged with our European colleagues to design a more sensible solution.
Some firms would like us to 'break glass' and implement a unilateral quick fix here in the UK. But I think that would be a mistake at this stage, not least because in the short-term there is a mechanism available to firms to mitigate the impact on business written before the introduction of Solvency II, namely the Transitional Measure on Technical Provisions, or TMTPs. We have allowed all firms that applied and qualified to use this measure. And data reported to us by firms indicates that the TMTP benefit across the life industry at the inception of Solvency II almost entirely offset the aggregate Risk Margin. I will also reiterate a point I've made previously, which is that when we consider whether or not firms are in a position to pay dividends, one of the main quantitative yardsticks we use is capital levels after the benefit of transitionals 2 . In other words, we regard transitionals as creating high quality capital.
Given the substantial benefit arising from TMTPs, it is important for both firms and the PRA that their level remains appropriate over time. And there is a mechanism to ensure just that, whereby firms may apply for a recalculation following a material change in their risk profile, or every two years. Our goal is that the maintenance of TMTPs should be as straightforward as possible and not onerous for any of those involved. The PRA recently closed a consultation 3 on proposals for this process and we expect to issue a supervisory statement on the subject later in 2017.
Another industry complaint is about the quantity of data firms must report, and the compressed timetables for doing so. It is a plain fact both that the volume of reported data is higher than under the previous regime and that regulatory returns are due to the PRA more quickly. I view both of these things as a significant improvement, which will enable better delivery of the PRA's objectives of policyholder protection and firm safety and soundness. To be frank, the situation we sometimes faced as regulators in the past, where firms were unable quickly to tell us their asset exposures, and solvency reviews sometimes occurred 18 months in arrears, was unacceptable, particularly given the fast pace at which serious economic and financial crises can develop in the modern interconnected world. However, I keep an open mind about whether we have struck the right balance - there may be a case for lowering the burden, if it turns out that some of the new reporting doesn't deliver supervisory benefits commensurate with its cost. We are also investigating the potential for publication of selected statistics to provide an overview for market participants.
2 http://www.bankofengland.co.uk/publications/Pages/speeches/2015/861.aspx
3 CP47/16, issued December 2016, consultation closed 15 March 2017,
http://www.bankofengland.co.uk/pra/Pages/publications/cp/2016/cp4716.aspx
4
On 16 November 2016, the Bank published its first edition of a new quarterly Statistical Release for the UK banking sector 4 . We are considering the potential for issuing something similar for the UK insurance sector, derived from the data insurers submit to us. We plan to engage with industry soon to discuss what format and content would be of most use and interest, with a first release potentially in the autumn.
A third topic of rumbling discontent relates to how easy insurers find it to invest in illiquid assets, and then to receive appropriate regulatory recognition. In my view, there are two quite separate issues at play here. The first is a certain 'clunkiness' in the design of the Solvency II Matching Adjustment (MA), a feature of the regime which gives insurers a capital benefit if they can match asset cashflows with liability cashflows. The rules were written on the assumption that insurers generally match annuities with simple securities, such as liquid corporate or sovereign bonds. This assumption was broadly true when the rules were agreed several years ago, but more recently insurers have increasingly turned their attention to other classes of investments that, generally speaking, have appropriate characteristics for meeting long-term insurance liabilities. But those types of investments often fail to meet certain rather prescriptive rules.
Some in the industry argue that we have the flexibility to ignore some elements of the rules. I don't think that is right. The directive is quite clear that cashflows from MA-eligible assets must be 'fixed', not 'predictable' or some other more flexible term. Now, it is not our job to unpick the law. But we should apply it sensibly. And I think there is plenty of evidence that that is what we are doing, whether in our approach to certain assets (for instance in allowing pairing of variable cashflow assets with derivatives, and restructuring of assets like equity release mortgages) or in the fact that we have allowed infrastructure assets in sectors like healthcare, social housing and telecoms infrastructure to achieve MA-eligibility. Now, given a completely free hand I would probably make modest design changes to the MA, perhaps with slightly greater flexibility around the definition of 'fixity' and a greater allowance for unexpected credit defaults. But I contend that we are implementing the regime we have today intelligently, with appropriate recognition of asset and liability matching leading to a £59bn reduction in capital requirements for the industry.
But the MA design itself is not the only reason that we at the PRA sometimes pause when firms discuss new investment opportunities with us. By their very nature, illiquid bespoke loans require very specific skills in terms of structuring, valuation and ongoing risk management. In particular, experience has taught us that the risk firms face from making such loans is related to how skilfully firms can manage any impairment process. These are not traditional areas of expertise for most UK insurers and, in my view, firms must clear a high bar in order to demonstrate that they have adequate systems and controls in place to manage these types of risks and exposures. We recently consulted on this topic 5 and look forward to further engagement with industry.
As a final word on this subject, when we have asked insurers directly about the impediments to investing in infrastructure assets, the number one problem raised is the lack of suitable investment opportunities, and not the regulatory framework. Perhaps they are just being polite but it seems to me that, when faced with the
4 http://www.bankofengland.co.uk/pra/Documents/regulatorydata/capital/2016/sep/bsrcrelease1609.pdf 5 CP48/16, http://www.bankofengland.co.uk/pra/Pages/publications/cp/2016/cp4816.aspx
5
cold reality of taking on these risks, insurers have often simply concluded that the yields available are insufficient compensation for the risks incurred.
## An over-zealous approach from the PRA?
I now turn to the other general theme from TSC submissions - whether or not we at the PRA have implemented Solvency II over-zealously. Or put another way, have we 'gold-plated' the directive? You will be unsurprised to hear that I say not. As I said earlier, it is only natural that there will be tension between the regulator and firms, because our objectives may be consistent but they are different. But rather than ask you to take my word on trust, let me provide some practical examples and attempt to explain how and why we have acted in the ways we have.
First, consider the challenge that the Association of British Insurers (ABI) made in its written evidence to the TSC 6 , that the Senior Insurance Managers Regime (SIMR) that we have introduced ' extensively 'gold plates' pillar 3 requirements of Solvency II '. The SIMR is designed to ensure the individual accountability of senior managers and directors of insurers for their own conduct, for overseeing the business conduct of key individuals reporting to them, and for the ongoing safety and soundness of their firms and the protection of their policyholders. As such, it is a key tool that the PRA uses to deliver its objectives. The SIMR is consistent with the provisions of the Bank of England and Financial Services Act 2016. So whilst the ABI is correct in its assertion that SIMR is not required by the Solvency II directive, it is the consequence of UK legislative choices and not simply an invention of the PRA.
Second, there is the thorny topic of approvals under Solvency II and, in particular, internal model approvals. The industry is quite right to point out that the overhead to gaining model approval is considerable. In some cases, weight rather than page length would have been a more appropriate measure of application size. But remember that firms with approved internal models in effect set their own regulatory capital requirements. Once we grant model approval, the PRA has only a very limited ability to adjust a firm's Solvency Capital Requirement. And as my colleague David Rule previously commented 7 'even acting in good faith, business pressure may create the risk that model changes generating lower capital requirements are favoured over time, and therefore that solvency standards might deteriorate' . In designing our approach, we were also mindful of the FSA's experience in approving internal models for banks under Basel II, where promised future improvements tended not to materialise. It should therefore come as a surprise to no-one that we approach model approval diligently, sceptically and rigorously.
Third, and relating to model approvals, is the issue of the PRA's own internal views on model calibration. We have previously commented on our use of quantitative indicators 8 , or QIs, but I think it is worth repeating that they are indicators and not pass/ fail standards. In particular, the standard for model approval is that firms must meet the tests and standards laid out under the Solvency II directive, and that is the basis for our
6 http://data.parliament.uk/writtenevidence/committeeevidence.svc/evidencedocument/treasury-committee/eu-insuranceregulation/written/47301.pdf
7 http://www.bankofengland.co.uk/publications/Pages/speeches/2016/921.aspx
8 SS17/16, http://www.bankofengland.co.uk/pra/Documents/publications/ss/2016/ss1716.pdf. CP20/16, http://www.bankofengland.co.uk/pra/Documents/publications/cp/2016/cp2016.pdf
6
assessments. We do not operate QIs as a minimum standard, but rather as an indication of the sort of calibration level we would expect to see for standard portfolios. Models falling outside of our QI ranges (both above and below) attract increased scrutiny.
Take as an example our QIs for corporate bonds. We feel that it is hard to argue that a model calibrated below the levels of credit spread increase we saw as recently as 2008-2009 meets the one-in-200-year standard required by the Directive. And so we set our QIs accordingly. But modelling credit risk is a complex business, subject to nuance and considerable expert judgement. Indeed, reasonable, knowledgeable experts might perfectly well disagree on some of the key assumptions. For those reasons, we have actually approved internal models with credit components that do not meet our QIs, where the firms in question demonstrated that their exposure is atypical and hence warrants a non-standard calibration, or provided off-setting conservatism in other aspects of the model.
## PRA's approach to insurance supervision
I'd like to turn now to some broader thoughts about what it means for the PRA to supervise insurance companies. I've previously spoken 9 about the history of insurance regulation in the UK and won't re-hash that here, since I've not yet reached the age where it is acceptable simply to repeat myself. Happily, according to our longevity QI I should have plenty of time for that further down the track. But I would like to provide some insights into why and how the PRA operates in the way it does.
The UK insurance industry is the fourth largest in the world. It employs 300,000 people in the UK, provides cover for nearly 27 million households 10 , and has £2.4 trillion 11 of assets.
It is therefore unsurprising that Parliament has given the PRA a General Objective of delivering safety and soundness of insurers, as well as banks. But for insurers only, Parliament has also given us an additional Insurance Objective - of equal weight to our General Objective - which is to contribute to securing an appropriate degree of policyholder protection. Now that sounds like rather a mouthful, but fortunately the Bank's Independent Evaluation Office (IEO) has today published a thorough evaluation of the PRA's approach to its Insurance Objective.
The Bank set up the IEO in 2014 to help Court evaluate the Bank's performance. Early last year, Court asked the IEO to consider whether we are doing enough of the right sorts of things to deliver the Insurance Objective. The IEO's report includes a number of interesting findings, which I welcome and find consistent with my own observations from my time as Executive Director of Insurance Supervision.
In particular, the IEO found that there is a lot of activity going on in pursuit of policyholder protection. This comes as no surprise to me. The PRA's insurance supervisors, actuaries and other professionals have in my
9 http://www.bankofengland.co.uk/publications/Documents/speeches/2016/speech877.pdf
10 Source: ABI factsheet https://www.abi.org.uk/~/media/Files/Documents/Publications/Public/2016/KeyFacts/ABI-
UKInsuranceAndLongTermSavings-KeyFacts.pdf
11
As at 2016Q2, based on insurers' reporting to the PRA
7
experience a strong public-service ethos. They are strongly motivated to protect insurance policyholders, which is for some a more human motivation than safety and soundness.
However, the IEO found that we could do more to explain what we mean by 'policyholder protection'. Some of this is straightforward, such as ensuring staff have clear legal analysis and understand where and when the PRA should and should not act, and we confirm in our response to the IEO's report that we will improve that analysis and training where required. But the IEO also raises some deeper questions about how much protection the PRA should seek to deliver, for which sorts of policyholders, and with what confidence. And it is those deeper questions that I intend to consider briefly in the remainder of my comments today.
Insurers are subject to what is known as 'twin peaks' regulation in the UK. In addition to supervision from us in the PRA, the Financial Conduct Authority also supervises insurers. And the IEO report makes it clear that, in the context of securing policyholder protection, the FCA has the primary role. The key word of the PRA's objective is 'contribute'. We should only act when there are prudential concerns, whereas the FCA acts on wider matters of fairness. To give an example: the PRA might intervene to prevent an insurer from selling new policies if it had inadequate financial resources, but it would be for the FCA to consider whether those policies contained unfair terms and conditions. In most circumstances the split of responsibilities is clear, but we need to be sure we are sufficiently explicit about the limits of the PRA's role in respect of areas such as with-profits business, corporate restructurings and firms that close to new business and go into run-off.
Another important consideration for us is whether the PRA should seek to protect all policyholders equally. The insurers we supervise received £260bn of gross written premium in 2016. Some might say that the PRA acts as though 'all policyholders are equal, but some are more equal than others'. But in my view there is no need to obfuscate in this manner and we should be quite clear that while the regulatory standard set in Solvency II is uniform, we do not in supervision attempt to cover all policyholders equally. It is in my view entirely appropriate that we should direct more of our resources towards policyholders who will suffer greater financial hardship if their policies do not pay out as promised, or those for whom there is the greatest asymmetry of information between company and customer, or who face the highest possibility of being unable to replace cover if it is lost.
For example, five million households have life insurance and two million have personal pensions. Some of the oldest and most vulnerable people in our society are reliant upon income provided by insurers under very long-term annuity contracts, into which they have invested their life savings. While the FSCS plays a vital role, the financial harm caused by the failure of an insurer in this space would be considerable and it is quite right that the PRA devotes significant resources to supervision of this sector, and that we have a very low appetite for any firm being unable to honour its promises.
In contrast, there is a range of other insurance contracts for which the consequences of firm failure are less severe, and hence with limited resource to spread around the PRA's supervisory approach can be somewhat less intense. For example, there are various classes of property insurance, such as mobile phone, for which an argument could be made that the PRA should restrict itself to a lower level of supervision. Our resourcing
8
model already works to some extent in this way, but the IEO has challenged us to think more deeply about this question.
Another interesting area is insurance bought by commercial or corporate customers. In these cases, the information asymmetry between buyer and seller is lower, or sometimes reversed, and contracts are typically renewable annually, meaning that there are lower concerns about customers being locked into long-term policies. In circumstances where buyers often choose their cover provider based on third-party measures of financial strength (such as opinions from rating agencies or insurance brokers), there is a case to be made that the need for the regulator is reduced. Indeed, many in the London Market have made this very point (somewhat less subtly!). But before anyone excitedly hears this as confirmation that the PRA will no longer regulate certain types of insurer, remember that we have a statutory objective to contribute to protecting policyholders of all stripes. So the debate is about 'how', not 'whether'.
In fact, the PRA's resourcing model already works to focus our resources on those areas of most concern. We categorise each regulated firm according to its size and the potential harm customers would suffer should it fail. The majority of firms we regulate are assigned the lowest category, reflecting the low level of risk they pose to our objectives. Firms in higher categories receive more supervisory and specialist resources, and decisions for those firms are approved at more senior levels within the PRA's decision making framework. The IEO have challenged us to look at this system again and we will do so. Our Executive Director for Insurance Supervision, David Rule, will take forward discussion of all of these issues with the Prudential Regulation Committee. We will then revise our Approach to Insurance Supervision document to explain any changes we decide to make.
In doing this we will of course need fully to factor in our third objective, to facilitate effective competition which was also recently reviewed by the IEO 12 . Although this is a secondary objective, it is given a high priority in the PRA and is also quite rightly a topic in which the Treasury Select Committee takes a strong interest. Our competition work spans banking and insurance, but there has been much more public debate about the former. We have, however, been pursuing our objective assiduously on the insurance side - to pick out some examples of steps we have taken to facilitate competition:
- we have created a very proportionate regime for the 40% of firms we regulate who are not covered by Solvency II;
- we have given around 85% of Solvency II firms in the UK the opportunity to avoid around 70% of the quarterly reporting burden they would otherwise face;
- we are working with HMT, HMRC and the London Market to introduce a new structure for insurance-linked securities; and
12 http://www.bankofengland.co.uk/about/Documents/ieo/evaluation0316.pdf
9
- we have authorised 19 new insurance companies and Lloyd's managing agents in the first three years of the PRA's existence.
## Conclusion
Financial regulation and supervision can sometimes be a dry business. And in insurance in particular, some of the more arcane aspects of the regulatory framework can rapidly turn off everyone except us and the industry.
I have therefore spent some time looking into individual policyholder cases, to remind myself about what is going on at the frontline. One in particular made a powerful impression on me. A young man got in touch with an insurance company to claim on his critical illness cover. The reason for the claim was that his 4-year old son had brain cancer. The cancer was in a difficult place, and the best-case outcome from a forthcoming operation was grim. The worst-case was the grimmest of all. The father was calling to claim on the policy in order to help cover the ancillary costs to the family while the NHS took forward this procedure. The company did the right thing, and it had enough money to do the right thing.
I and others were only exposed to this case in a wholly anonymised way, for data protection reasons. Despite this I have never been in a meeting where so many people wept.
This is what policyholder protection is really about. It's not about the Risk Margin, the view to ultimate, the Matching Adjustment or corporate bond spreads. It's about making sure that insurance companies are there to help when disaster strikes. That's why we take this part of our job so seriously.
10