Solvency 2
Speech by Sarah Wilson, Director Retail Firms Division, FSA
City and Financial Conference
20 March 2007
Thank you for the invitation to give your key note speech today. I am delighted to be here.
As I know you will agree, Solvency 2 is of huge significance for the insurance industry – not only because it will set prudential standards within Europe that are unlikely to be revised further for some considerable time, but because the approach adopted may well influence global standards more widely. In this context, my agenda today is fourfold – to set out our high level objectives for the project; to comment on progress and outstanding issues as we see them; to note the significance of further quantitative work by the industry in resolving those issues; and to look forward to the implementation challenges.
High Level Objectives for Solvency 2
So first our high level objectives. Our approach to Solvency 2 has been informed by the same principles that have underpinned our domestic reforms. The prime motivation, underpinning all our more specific objectives, has been to increase focus on risk management, through improved risk measurement and a link to capital planning - and to provide appropriate incentives within the new framework to do all these well. After all, fundamentally the business of insurance is about informed and controlled risk-taking, and regulation should respond to that.
More specifically, what are we trying to achieve in our negotiations on Solvency 2? We have approached the project with a number of more detailed objectives in mind.
First, we have looked for an appropriate adaptation of the Basel three pillar framework. This involves incorporating in Solvency 2 a ‘Pillar 2’ supervisory review process to heighten focus on risk management – very familiar to UK firms from our domestic reforms. And it involves introduction of improved disclosure through ‘Pillar 3’, so that market discipline complements regulation.
Second, we have argued strongly for market-consistent valuation standards, including in assessing the scale of liabilities to policyholders.
Third, we believe capital requirements must reflect risk in both assets and liabilities (including any interactions between the two). To do this properly, they must reward real risk diversification; and they must take account of the extent to which risk-transfer instruments mitigate and transform risks that a firm retains.
Fourth, firms should be allowed to use their own internal models to determine their regulatory capital requirements, subject to appropriate controls over the adequacy of those models.
Fifth, we need a framework that better caters for groups. This is of major significance - 50% of European direct insurance is written by the 20 largest groups, and they operate in a number of Member States through different subsidiaries. These groups currently face overlapping and varying requirements from an array of home (parent) and solo (subsidiary) supervisors. Furthermore, policyholder protection is hampered by the absence of a framework that facilitates regulatory understanding of the group as a whole as opposed only to the group simply as the sum of its parts. In other words, understanding of group capital, group risks and group controls should be an important additional part of the European supervisory framework going forward, with an associated clear allocation of regulatory powers and responsibilities.
Finally, we have looked for appropriate ‘harmonisation’. Convergence in regulatory requirements should be welcome news for both the industry and consumers – it promotes competition and aids transparency. Given the fragmented nature of the current European insurance regulatory framework, with varying national standards and approaches, a common set of objectives and conceptual approach will be a major step forward. And it will be a good springboard from which to make further progress in achieving real convergence in supervisory practice. But it is also important to remember that if harmonisation of objectives and concepts turns into a search for harmonisation at every level of detail we risk loosing a risk sensitive regime – the capital requirements for each firm may no longer reflect the local or sectoral markets in which that firm operates. Where that happens harmonisation could result in greater inequality and increase the potential for perverse or pro-cyclical effects. Achieving the appropriate (most risk sensitive) level of harmonisation involves a difficult balance between common safety standards and principles and flexibility to reflect the risks in local markets.
Solvency 2 Progress and Issues
Let me turn then to our view of progress as against these objectives. It is worth stressing of course that the final framework will be the result of negotiations. CEIOPS (the Committee of European Insurance and Occupational Pensions Supervisors) has provided technical advice on the drafting of the Solvency 2 to the Commission, and has been the forum to date for forging a common view amongst EU supervisors. But clearly, negotiations proper will commence in the Council of Ministers and the European Parliament, when the Commission’s Proposal for a Framework Directive proposal emerges. At that point, the FSA hands over the UK negotiation baton to HM Treasury. And that stage of the process will clearly be subject to its own dynamics and attendant uncertainties. Based on the Commission’s stated timetable, we still expect the Directive Proposal to emerge in July.
But, standing where we are today, based on negotiations in CEIOPS, and what we have seen of draft directive text, what is our assessment of the progress, and what are the outstanding issues?
Overall, we are encouraged by progress in CEIOPS. Perhaps the single most important aspect has been agreement on a market-consistent valuation standard for liabilities, and in that context, on the cost of capital approach to setting a risk margin over best-estimate for non-hedgeable risks. In addition, we consider that there has been good progress on the form of a standard formula for calculation of the risk based capital requirement (the SCR), though naturally there is ongoing work to refine the calibration. Also important has been establishing a good foundation for a principles-based approach to the approval of use of firms' own internal models for regulatory capital purposes.
In terms of open issues, the matter of the calibration and form of the minimum capital requirement (MCR) formula clearly remains to be resolved, and will be very important to the coherence of the whole framework. In addition, the definition of own funds (or available capital) still remains relatively open, with the added complication of a timing mismatch between Solvency 2 and the own funds review for banks. And last, but by no means least, discussions are very much ongoing on an appropriate supervisory model for groups.
Further quantitative analysis
Overall then, much has been achieved but there are some remaining hurdles to jump. And, while negotiations continue, we are also very keen to continue to bring all available evidence to bear on the debate – to maximize the chances of a regime that is genuinely risk sensitive and robust. Consequently, I would like to take this opportunity to emphasise to firms the benefits of participating in CEIOPS third quantitative impact study. Of course, participation in QIS3 is voluntary. But I would strongly encourage you to take part if at all possible, and for two reasons.
First, the results of QIS 3 will inform negotiations and influence the final form of the new regime. Decision-makers need data to make sensible choices between options. The results of QIS 3 will coincide perfectly with the start of negotiations. It does not matter that they will follow publication of the Framework Directive in July - if the results indicate that there are big problems, there will still be time to change the framework.
Second, participating in QIS will help you understand the likely impact of Solvency 2 on your firm and will help you plan for implementation of the directive. It is a very practical way to get Solvency 2 on your Board's radar.
I understand that some firms might conclude that, after two previous quantitative studies already, there is little further to be gained by participation in the third one. This is however absolutely not the case. QIS 2, run at roughly the same time last year, was about design of the Pillar 1 framework, particularly in respect of solo entities. QIS 3 might then be characterized as being more about calibration, to further the work on the more detailed implementing measures that will underpin directive text, and on which CEIOPS is already starting to work. But that is not the whole story. Some important aspects of the Solvency 2 framework were not tested in QIS 2, and will be tested in QIS 3 for the first time. These aspects should inform the shape of the directive itself that is being negotiated as the QIS results emerge.
Most specifically, and beyond helping in the calibration on technical provisions and the SCR where conceptually most progress has already been made in CEIOPS, QIS3 will especially help to influence debate on the three remaining open issues mentioned earlier – the form of the MCR; the appropriate supervisory model for groups; and the definition of Own Funds.
On the MCR, QIS 2 revealed that in too many instances the proposed calculation approach gave a result was too high compared to the SCR, and in a number of cases, in particular for some life firms, it was actually higher than the SCR. This clearly has a profound effect on the risk sensitivity of the regime overall and of supervisors’ ability to intervene in a sensibly gradated fashion. Finding an appropriate MCR will be highly important to the success of Solvency 2. QIS 3 will test a proposed MCR calibration and two formulae intended to deliver that calibration – one currently preferred by CEIOPS and one championed by parts of the industry. Both the calibration and the form of the calculation remain open issues. QIS 3 may be the best chance the industry has to influence the outcome of that debate.
On groups, QIS 3 will be the first opportunity to test the measurement and potential size of diversification benefits at the group level, as well as the extent to which surplus capital is transferable between entities to meet overall group capital requirements. This is key to understanding the need for and viability of a new approach to group supervision.
On capital, QIS3 will test the high level own funds definition that the Commission has set out in the draft text.
Finally, some firms, I know, will consider that they do not have enough resources to commit to a complete execution of all aspects of QIS 3. To those firms I would say that it is still important that you participate, even on a partial basis. Partial results may still yield useful information and are certainly better than none. In particular, I would advise you to concentrate on those aspects which are the most material for you. While this approach will not tell you (or us) the full impact of Solvency 2 on your firm, it should yield useful information on the suitability and practicality of the proposed calculations. To larger firms, and niche players, that may have internal models, I would like to encourage you to submit appropriately calibrated internal models results, along-side standard formula numbers – especially for underwriting risks where there is little available data. Standard formula results give us certain information about impact, but tell us little about the accuracy of calibration.
We are aware that QIS represents a big resource commitment for firms, and we are appreciative that firms are prepared to commit resources. We, too, have committed material resources to it. You are invited to contact our policy experts direct – QIS3@fsa.gov.uk. And I would commend to all of you the QIS workshops that the trade associations and FSA are running jointly.
Looking forward to implementation: challenges for larger and smaller firms
My last set of remarks relates to implementation. It may seem strange to consider this at the current stage - before the directive text has emerged, let alone been finalized. But the Commission has an ambitious timetable and, in practice, implementation may only be four years away.
Firms should not fall into the trap of thinking that Solvency 2 is ICAS. Of course, firms' ICAS work is directionally correct in that it is founded on the same principles that underpin Solvency 2, and it has given firms a head start in the development of modelling techniques. But in order for UK firms to maintain that advantage, they will need to keep improving and should not overlook the recent strides that firms in a number of other Member States have recently also made in this area.
We consider that Solvency 2 will present rather different implementation challenges to different populations of firms. Larger firms and others that may be planning to use their internal models to calculate all or part of their SCR will need to plan ahead to achieve the necessary regulatory approval. And this may well necessitate an improvement in the standards that firms currently attain. Many smaller firms, on the other hand, may have no such ambitions, at least in the shorter term. But Solvency 2 will still present challenges. Let me talk a little more about each group of firms in turn:
Internal models
In the UK context, ICAS is part of Pillar 2, and the current directive and realistic reporting requirements provide the Pillar 1 underpin. In the Solvency 2 context, regulatory approval will permit firms to use their internal models to generate their Pillar 1 capital requirements (the SCR). While Pillar 1 also has an MCR, this is by definition a level of capital which represents an unacceptable risk to policyholders, and at which a firm may expect to be placed into run-off. So, for determining the appropriate operating level of capital for the firm, there will be a high level of reliance on the internal model number, and as a result, firms may expect particularly robust regulatory challenge before regulatory approval is granted.
CEIOPS and the Commission have made good progress in developing a principles-based approach, which allows firms a relatively high level of modeling freedom, and regulatory approval will hinge on a firm satisfying three broad tests: a statistical quality test; a calibration test; and a 'use test'.
It is not my intention today to give a workshop on internal model approval, but nevertheless let me offer firms a few initial pointers by unpacking a little more what those rather catchy titles will mean in practice. First, and perhaps most straightforwardly: the calibration test. Happily, Solvency 2 calibration is essentially in line with ICAS, and therefore firms may look directly to their ICAS experience, and especially their ICG, to assess whether they are likely to meet the calibration standard.
That leaves the first and third tests, which are the areas where firms are most likely to need to build on and further improve their current ICAS practices. The use test is essentially about the relevance of the model to how a firm runs its business – we must not forget that one of the prime reasons for permitting models to be used for regulatory purposes is to incentivise better risk management and allow a regulatory dialogue based on a firm's own assessment of its risk. So, we will expect that the regulatory capital number will be generated from the model that is actively used in the management of the business. For this to be so, we will be looking to see that the model is firmly 'embedded' – that is, that it plays an integral role in risk management, decision-making and capital allocation processes.
The statistical test is essentially about the predictive ability of the model. Completeness and accuracy of input data, modeling methods and assumptions, and methods of aggregation are all relevant here. The regulatory approval process will focus particularly on a firm's governance and controls surrounding the model. Firms will be expected to undertake a regular cycle of model validation that includes monitoring the performance of the model, reviewing the ongoing appropriateness of its specification, and testing its outputs against outcomes. Firms will be expected to be able to attribute experienced profits and losses to each major business unit and, in addition, to demonstrate that the risk drivers and categorizations in the model align sufficiently with the causes and sources of profits and losses. Of course, model validation will also be concerned with the stability of the model and in this area it will be relevant to test the sensitivity of model outputs to changes in key underlying assumptions, including, naturally, aggregation assumptions.
I'll round up on internal models by saying that gaining internal model approval will be an intensive process, but one with which the FSA is very happy to engage actively in the run up to Solvency 2 implementation in response to any industry demand.
Standard approach
For smaller firms the Solvency 2 challenge is likely to arise principally from the need to be able to produce a robust best-estimate of their liabilities. While Solvency 2 may deliver relatively simple methods to calculate a market-risk margin in technical provisions, and to calculate the capital requirement, these numbers are, when all is said and done, rather smaller than, and in any case dependent on, the best-estimate. In principle, it is not unreasonable to expect any insurer to have a robust understanding of its best-estimate, since that is the bread and butter of its business. Smaller firms, which may not have a rich data histories, face a particular challenge, though. And, in planning for implementation, we see two complementary ways forward. First, as CEIOPS develops implementing measures to support the directive text, we hope there will be an accompanying workstream for actuarial standard setters to develop clear actuarial standards that fit within this framework. (The aim would be to define, limit and describe appropriate deterministic methods to be used when stochastic methods are not appropriate; and to determine how they should be calibrated and how and when they should be used.) Second, industry trade associations may have a useful role in helping firms to broaden their access to data, by collecting and collating data from individual firms and so giving firms access to 'industry data'.
We have made these suggestions before, and while we have seen some signs of emergent activity in relation to actuarial methods, there is clearly further to go, as CEIOPS' work progresses. In relation to data, we have not so far become aware of any such initiatives, and we would encourage the industry to examine more actively what might be achieved.
That concludes my remarks for today. I hope you will agree that Solvency 2 has come a long way since the idea was conceived. But what is equally clear is that there is still some way to go. As against our priorities, there is much in the work to date to be pleased about, but there are also important debates remaining and ones that the industry can and should help to influence through participation in QIS3. Looking further ahead, firms need – whether or not they plan to use internal models – to start preparing for implementation, remembering that while ICAS has been useful preparation, it is not the same as Solvency 2. Indeed, one lesson we have all learned from the implementation of Capital Requirements Directive in the banking sector, is that the implementation challenge should never be under-estimated!