Speech by Paul Sharma, Head of Prudential Standards, FSA
AISAM / ACME Joint Conference, Bruges
19 October 2006

Madame Beres, Commissioner, Ladies and Gentlemen

I would like to thank AISAM & ACME for having given me this opportunity to address you today, and to do so in such a distinguished company as Madame Beres and Commissioner McCreevy – as a humble technician and an adviser to the truly influential parties that will eventually decide on Solvency 2 - I am honoured.

As the Chair of CEIOPS Solvency 2 Pillar 1 Working Group I have had, and hope to continue to have, close contact with the representatives from the mutual sector. It is therefore only natural for me to want to outline my views on how I see, after over 2 years of preliminary discussions, technical studies and consultation, some of the key questions of the Solvency 2 framework can be resolved.

I would like to start by reiterating what I think Solvency 2 should achieve – namely – to create a regulatory environment which incentivises and rewards insurance firms to use modern risk management practices that are appropriate to the size and nature of their business.

In one of the early Solvency 2 background reports which analysed insurance failures and near misses and which carries my name – the Sharma Report – we found that the main causes were clustered around the broad themes of management quality and inappropriate risk decisions rather than inadequate capitalisation per se. So, it is quite clear to me that all aspects of the framework – Pillars 1, 2 and 3 – should be geared towards improving and incentivising firms to use modern risk measurement and management techniques.

Clearly, fostering robust risk management in insurance companies is only one of the building blocks for a successful prudential framework - protecting policyholders and maintaining market confidence also require adequate levels of capital. The focus of my Pillar 1 working group has been to flesh out what this should mean in practice - in essence - to create a more risk-sensitive and risk-responsive capital requirement that not only takes account of the risks on the liability side, but also on the asset side, and gives due credit to the use of risk mitigation techniques.

In my mind, and I am encouraged to say that now the majority of my CEIOPS colleagues would seem to share this view, one only knows what is and what is not adequate, by first measuring and valuing assets, liabilities and risks as they really are. Economic reality must be the starting point. If I were to be asked to name the most important principle that Pillar 1 should adhere to, I would without hesitation choose this one.

Which leads me to commenting on the vexed question of valuing liabilities and
technical provisions. This is an area where Solvency 2 in particular should follow the principle I just outlined above. If liabilities are set using unharmonised margins (as is the case currently), there is a risk that they will be inadequate or excessive for their purpose, either putting policyholder security at risk or imposing excessive cost on policyholders.

The purpose of technical provisions is to protect policyholders even if all of the capital is used up – that is to say - even if the 1 in 200 loss which is the Pillar 1 prudence objective for the Solvency Capital Requirement has occurred. Technical provisions best achieve this by including a sufficient margin in excess of the best estimate of policyholder liabilities to make it attractive for investors either to recapitalise the firm back to the level of the SCR or, where possible, to transfer the liabilities to another insurance company capitalised to at least to the amount of the SCR. The margin therefore should reflect the price to be paid for reacquiring capital equal to the amount of the SCR and therefore needs to be set equal to the cost of that capital.

The forthcoming CEIOPS Pillar 1 CP, which I strongly urge you to study it carefully and respond, will further discuss the issue of technical provisions and the approach to setting a risk margin. Although this Consultation Paper is still subject to a final decision by CEIOPS Members' Meeting in Budapest, I can say that a great number of supervisors have accepted the validity and feasibility of the Cost of Capital method, but that for long-tail non-life insurance classes further work and consideration is still needed.

I believe the Cost of Capital approach is conceptually better and moves us much closer to the ideal of market consistency; and second, for consistency and simplicity of the framework, I see no reason why the Cost of Capital approach could not be used by all types of firms – large and small - particularly as the prescribed, simplified Cost of Capital method can be implemented with relative ease.

Turning now to the standard approaches, something that I know is certainly of interest to you all. It goes without saying that getting Pillar 1 right is of central importance. This is particularly so if we want Pillar 2 add-ons to be used less frequently, as has been indicated by the Commission. But, as the standard approaches will need to be applied by a great number of different firms with varying risk profiles, it is not an easy task to come up with a single, uniform standard formula that delivers a reflective capital requirement which is rarely subject to add-ons, and which gives harmonised outcomes.

In the Pillar 1 working group where there is much lively discussion over this issue, further work is still needed to achieve an agreement between those who favour the use of more factors than scenarios and vice-versa. I see that a possible compromise capable of commanding wide support across all jurisdictions might be that within the standard formula, for some risk modules, there might be more than one calculation method.

Based on some objective criteria, the "basic approach" calculation method in the standard formula could be substituted by some other harmonised calculation method, so as to better reflect the specificities of that particular firm but nevertheless arriving at the agreed common overall level of safety. So, in practical terms, it could be envisaged that for market, life and non-life cat risk modules in the standard formula there should be a "basic approach" and a more "advanced approach" that uses the outcome of stress and scenario tests or factors that are tailored to the circumstances of individual firms. From a technical point of view, this approach is particularly useful in cases where the risks which a firm faces are not the simple linear risks that are best suited to measurement by a simple risk factor-based formula. And, by adopting this approach we will be able to accommodate the great variety of firms in the EU, particularly those who underwrite very specialised lines of business, as well as to encourage and incentivise firms to improve their risk management. Furthermore, this should smoother the transition in firms to the use of internal models.

Clearly, whatever methods are chosen they must be inter-changeable, that is to say, both the simpler factors and the scenarios should capable of delivering broadly equivalent outcomes. This is one of the key questions that we will need to resolve for QIS3 and on which I would welcome further views, thoughts and evidence from the industry.

I would just like to note that we do not see by any means that the use of stress and scenario tests are beyond the mutual sector's reach. Our experience in the UK following the introduction of our ICAS framework would suggest that with the right kind of guidance and help, also the smaller firms have been able to make use of stresses and scenarios in their individual capital assessments, and perhaps more importantly, in the process the firms' senior management (and their supervisors) have learnt a great deal about the firms' financial health and risk profile. So, from the point of view of strengthening the prudential framework, I do see part of the future being a much wider use of, and reliance on, stresses and scenarios by firms of all sizes.

I would now like to mention a few words about the considerable challenge that the policy-makers need to deal with when deciding on the new framework – namely – how will we be able to set things in stone now, when it will actually be implemented some time after 2010 / 2011? Between now and then and whilst the cogs of the EU legislative processes go round, best practice in the insurance sector will have moved on, as will have information technology capabilities, modelling approaches, risk management techniques and maybe the financial markets as a whole.

The response to this challenge is by no means easy but, essentially, I believe it requires policy-makers to be both visionary and courageous.

Visionary, as we will need to have it clear in our minds where we would like the insurance sector to be in terms of standards for risk management in 10 years from now; and courageous, as we will need to avoid setting policy based on current practices but rather base it on a balance vision of what the future should be like.

It is now widely accepted that the direction in the industry is towards more focus on risk management and modelling, and so Solvency 2 should further encourage this as well as seek to capture the benefits from this move for supervisory purposes.

How should insurance firms approach Solvency 2? Having seen the flurry of activity around the implementation of MiFID and CRD and the other Financial Services Action Plan measures, I think insurance firms are well placed to learn from those experiences. So speak to your colleagues in the wider financial services industry – they have first hand knowledge of how to approach the implementation of major pieces of EU legislation.

The key to firms' successful response is that firms should not try to ignore this profound change but start engaging and preparing for it now. Change can be painful. But the pain, when distributed across a number of years, will be much more bearable. Participating in the quantitative impact studies, although time and resource consuming is a good way of getting an early feel as to the changes that are coming your way.

Being a chair has meant that I have received numerous representations from the industry. I have been struck by the impression that the industry by and large is very supportive of this project – and I believe this certainly must be an extreme tail-event, 1 in 1000 years in the field of financial services regulation! – and that the industry more generally is embracing modern risk management techniques. I also get the sense that the industry recognises that, whilst to begin with new approaches will demand more from the industry, the benefits that will flow from a modern and flexible risk-based framework will outweigh the trouble of changing current practices. I believe this attitude is very healthy in ensuring the long-term prospects and global competitiveness of the European industry.

Although I would like to think that CEIOPS and its Pillar 1 working group have made substantial progress over these 2 years, I know that you sometimes feel that we supervisors are not as willing to embrace change as you, the industry, would like us to be. This is clearly an area where you still have lot of work to do – convincing your supervisors (and finance ministries) that a truly risk-based and market consistent Solvency 2 is what you actually want!

If Solvency 2 is a challenge for firms, it is also a challenge for supervisors. So, it does appear to me that if the policy-makers have the right to ask the industry to up their game – this also applies in reverse. We, the supervisors, need to be honest and admit that we need to shift the skill sets of our staff from understanding rules to understanding markets. This reflects the change of the supervisory regime from a rigid tick-box compliance monitoring approach to a more flexible and interactive dialogue and supervisory review process with firms.

Before concluding my remarks I would like to say a few words about the impending 3rd Quantitative Impact Study. Thank you all those mutuals that participated in QIS2. Whereas QIS2 mainly focussed on testing the structure and design, QIS3 will be our main vehicle to test and get the calibration of the standard formulas right. Much work still needs to be done between now and early next year so that we are well placed to put considered and well-supported proposals for the calibration forward.

All the quantitative data, as well as the qualitative input from QIS2, is of immense value, but we will undoubtedly need further input and evidence from you before we are in a position to launch QIS3. I shall not make any apologies for this as I know it is in the best interests of both the industry and the supervisors that we get the calibration right. I cannot emphasise enough the importance of having as many companies as possible participating in this exercise. Evidence-based policy-making is worth the effort.

One of the famous sons of Bruges was Simon Stevin, who, among other things, was a famous mathematician and an engineer of his time. One of his achievements was that whilst a quartermaster in the Dutch Army, he invented a way of flooding the lowlands in the path of an invading army by opening selected sluices in dykes.

If I am correct in suspecting that there might be a number of you in the audience who are looking at Solvency 2 and seeing the floodgates starting to open already, I say I have sympathy with you. Change and uncertainty can indeed be unnerving.

Much of the preparatory work has been done and soon it will be up to the law-makers – Madame Beres and her colleagues in the European Parliament, Commissioner McCreevy and the Member States – to take the key political decisions. By engaging early and being active participant in the debate and the QIS, I am confident that the mutual sector, as well as the insurance industry as a whole, will be able to withstand the rising tide, and ride the high-waves of Solvency 2.

More Speeches: