Edinburgh, 26 February 2003
John Tiner
Managing Director, Financial Services Authority
Realistic Liabilities, With-Profits Governance and the Role of Life Actuaries

First let me express my thanks to you for inviting me to speak today. This conference is billed as the ‘highlights’ of the November 2002 Convention for Life Actuaries at which I spoke, but events have since proceeded at such a pace that I feel it is appropriate to deliver a new speech that largely concentrates on subsequent developments. As you know earlier this month I took an important initiative in advancing the reforms that we were planning, to place in increasing emphasis on a life insurer’s realistic financial position in our prudential regulation. I shall use my speech to explain this initiative in more detail. I shall also explain how this initiative interacts with our other major initiative on with-profits governance and the role of life actuaries. I gave an outline of our proposals on governance and the role of life actuaries in my original speech in November. This was followed by our consultation paper ("CP 167"), which was in preparation at the time. Both of these pre-dated our initiative this month on realistic liabilities which has important implications for the timing and, to some extent, the substance of what we propose on governance and the role of life actuaries.

As you will no doubt know, I wrote to the CEO’s of the major life insurers earlier this month. My letter referred to the significant impact that falling equity prices were having on the financial position of life insurance companies and to the difficult decisions that Boards and senior management of those companies were having to make about what actions to take to protect policyholders. It reminded them that we stand ready to discuss these decisions with individual firms and, most importantly, set out particular considerations that we would take into account in those discussions. In particular the letter referred to our previously stated intention to move from the current approach to a new, realistic approach and announced two important new steps in that respect.

  • First, we now plan to introduce this realistic approach from early 2004. We will also consider making rule changes to facilitate an earlier transition to this new approach. For example we are looking at, perhaps, allowing or requiring life insurers to report their end-2003 position on both a realistic and statutory basis in their March 2004 public regulatory returns. We are also considering whether and what transitional provisions, exemptions or special provisions are needed for smaller life officers.
  • Second, in the meantime, the letter invites life insurers to apply to us for a waiver of, or modification to, existing rules to accelerate some or all of these changes in their particular case. I shall now explain in more detail whether, when and how we would be willing to grant such rule waivers or modifications.

The starting point for this explanation is to emphasise that our prime consideration in deciding to waive or modify rules is the protection of consumers. However for with-profits policyholders there are two aspects to this protection.

First, life insurers need to have financial resources that are sufficient to a high degree of confidence to meet guaranteed contractual liabilities. Second, with-profits life insurers need to pursue an investment and business strategy that enables them to declare and pay future discretionary bonuses that are fair and reasonable. It is important therefore that regulatory financial-resources rules are sufficiently prudent to secure contractual liabilities but not excessively prudent so that they unnecessarily restrict investment freedom and so prejudice a life insurers ability to declare future discretionary bonuses. This leads to two consequences.

  • First, waivers and modifications of rules – at least of the kind contemplated in the CEO letter – will only be granted to those life insurers that are financially strong as measured on a realistic basis. That is they must have financial resources at least equal to their realistic liabilities plus a safety margin. I will explain in a few minutes how this safety margin is to be determined.
  • Second, we contemplate granting the waivers and modifications of rules referred to in the CEO letter in a wide range of circumstances and not merely where a firm finds itself pressing against its statutory required minimum margin ("RMM"). The waivers or modifications target the harmful consequences that arise where the existing rules are excessively prudent as compared to the realistic basis. These consequences become most acute when a life insurer that is financial strong on a realistic basis finds itself pressing against its RMM but they are not confined to that situation. They arise wherever a life insurer finds itself unnecessarily restricted from pursuing an appropriate investment strategy, e.g. holding or buying equities where it reasonably considers this to be prudent and good value for policyholders in the longer term. Our view that we will contemplate granting the waivers or modifications in a wide range of circumstances reflects our longer-term aim to amend our rules for early 2004 for all major firms to remove the need for individual waivers.

I will now set out in a little more detail some of the important practical and actuarial aspects of how we will waive or modify rules. This explanation anticipates a letter to life insurers that we propose to send out next week. That letter will state that in essence a life insurer that requests a waiver or modification must do three things. Before I talk about these 3 things though, I should perhaps, remind you that the tests that must be satisfied according to our rules before a waiver can be granted are that:

  • The rule is unduly burdensome; or
  • The rule doesn’t meet the purpose for which it was made; and
  • It must not put consumers at risk

Clearly, in judging whether these tests are met in individual firm circumstances, we will require firms to submit their case to us and our analysis and evaluation of this will inform our decision. So this brings me back to the 3 things that will be included in our letter next week.

  • First the life insurer must calculate its realistic liabilities and the safety margin needed in excess of those liabilities.
  • Second it must compare its realistic liabilities plus safety margin to the realistic, i.e. market, value of its assets to determine its realistic position and so to demonstrate that it is strong on a realistic basis.
  • Third it must identify the particular rule waiver or modification for which it wishes to apply.

I will explain each of these three things in turn. First as I have said a life insurer must calculate its realistic liabilities. To form a satisfactory basis upon which the FSA could grant a rule waiver or modification this calculation must be robust and reliable. I will set out in a moment the principles that we would expect a typical life insurer that uses an asset-share methodology to apply in calculating its realistic position. However before doing so I would emphasise that

as practice is still developing in this area we do not expect life insurers necessarily to meet these principles in full, but we do expect them to have made significant progress toward implementing these principles and to have committed themselves to a clear plan to ensure that continued progress is made. For the same reason we also recognise that life insurers may wish to apply alternative, arguably better, principles to achieve the same objective, that is a robust and reliable calculation of the realistic position. We are open to persuasion that alternative principles are better either in particular cases or in general. With those caveats I will now set out the principles we are minded to follow.

  • The first principle is that there should be a clearly documented methodology for calculating asset shares, smoothing costs, GAR costs and other guarantee costs. This methodology should be consistent with the life insurer’s actual practice in setting bonus levels and with how that practice has been described to policyholders (e.g. in sales literature). For asset shares the methodology should state, and justify, which items are included, for example describing whether, how and why difficult items such as costs of guarantees, costs of smoothing and miscellaneous profits and losses from lapses or surrenders are, or are not, to be attributed to asset shares. Where individual asset shares are not separately calculated, the methodology should describe the statistical methods used to calculate the aggregate asset share. For smoothing costs the methodology should itself be based on a clear and documented bonus smoothing policy that is understood and approved not only by the actuaries who carry out the calculation but also by senior management. For GAR and other guarantee costs the methodology should deal with issues such as how take up or persistency rates are estimated and how the time value of options is taken into account. For options we prefer market value where the option is fully hedge with the purchase of suitable ‘swaption’ assets or the use of stochastic modelling where the option is not hedged. However, we recognise that techniques and data are not always available to allow stochastic modelling methods to be used and so, in the shorter term, we would also accept a suitably prudent deterministic method. I am aware that methodologies for the stochastic assessment of the value of guarantees are being developed by actuaries and other professionals. These would combine an appropriate stochastic model for projecting the value of equities and other investments, and appropriate time horizons and risk of ruin, along with a dynamic investment and bonus policy as agreed by the firm. In the longer term, this may well lead to an approach that is broadly consistent across the different financial sectors. However, there is still a considerable amount of work to be done in fully developing this methodology, choosing and validating appropriate investment models, and applying such a framework in practice.
  • The second principle is that adequate systems and procedures should be in place to put the methodology into practice. This requires adequate systems, data and controls. It would be understandable if Management thought they required some assurance about their firms systems and controls necessary to support the realistic calculation and perhaps even some comfort on the calculation itself, in which case firms may wish to consider putting in place an independent review of both the appropriateness of its methodology and the adequacy of its implementation. This might take the form, for example, of a formal or informal review by consulting actuaries and/or the life insurer’s auditors. I shall say more on this in the second half of my speech.
  • The third principle is that in order to show that it is strong on a realistic basis a life insurer needs to demonstrate not only that it has sufficient financial resources to meet its realistic liabilities under normal circumstances but also that this would remain the case in stressed circumstances. This is why a safety margin is needed in excess of normal realistic liabilities. There are a variety of ways in which this safety margin might be determined. At its simplest, we might apply a series of static stress tests focusing on key risks. A more sophisticated approach might be to use dynamic or stochastic financial modelling techniques. Whichever approach is used the key design issues include which types of risk to stress test or model, to what quantum or confidence level and over what time horizon, as well as how the results should be combined – that is should credit be given for supposed favourable risk diversification. At this early stage a more simple and achievable approach is needed. So for rule waiver or modification applications we are proposing a static stress test. In the very short term (ie. next month or two) this is likely to be based solely on an equity value stock. In the medium term, when sufficient data becomes available, we may decide to base this on a wider range of issues including key market risks such as equity value, property value and interest rate shocks and perhaps also on the key credit risks in the asset portfolio. We are undertaking work to set the quantum of these shocks. For this exercise the time horizon assumes the shocks are sudden and immediate. When, and if, the shock actually occurs we would not contemplate a further shock – at least not one on anything like the same scale. We also need to consider whether directly to aggregate the results of these shocks or to absorb their effect by recognising favourable risk diversification, but we have yet to reach a firm view on this. Restricting the calculation only to a few simple market and credit risk shocks and aggregating the results keeps the calculation simple and arguably any over or under prudence in the different aspects of this approach roughly offset. Although simple and static in its basic form to make it accessible at short notice to a wide range of life insurers there is one aspect of this calculation for which we would like a more dynamic approach to be taken. The stress test asks how the realistic value of asset and liabilities would change following a market and credit risk shock. The calculation of realistic liabilities refers to asset shares, cost of smoothing and the value of guarantees. The latter two depend in particular on how respectively the life insurer itself and its policyholders would behave in response to the market and credit risk shock. The life insurer needs to understand both how it would change its smoothing policy in response to the change in asset share values and how policyholder take up rates for options and guarantees would change as a result of those options or guarantees moving into (or deeper) into the money.

The three principles I have just set out refer to our approach to granting waivers in the short term. I have already mentioned that from early 2004 we intend to amend our rules directly dispensing with the need for individual waivers. We propose to develop these rules in a way that is consistent with the principles that apply in the short term for waivers. However we will also seek to make further progress especially in respect of the way in which the safety margin is calculated and on whether and how an independent review is implemented.

So far I have described how we would expect a life insurer to demonstrate that it is financially strong on a realistic basis. The next step is to compare a life insurer’s realistic position to its regulatory position. This is important because we are not proposing dispensing in toto with the existing regulatory regime which prescribes rules for the calculation of mathematical reserves, other liabilities and the required minimum margin. Waiving this in toto would not be permitted under the EC directives on insurance regulation. The EC directives lay down some detailed rules on how mathematical reserves, other liabilities and the required minimum margin should be calculated. Our UK rules implement these EC rules and add more detail. It is only in the area of this extra detail that we would grant rule waivers or modifications and this mainly relates to the mathematical reserves. However we believe that, for the present circumstances, this gives ample scope to modify the way in which the regulatory rules work to bring them closer to the elasticity of the realistic basis. The mathematical reserves after the impact of any waiver would not necessarily equal realistic reserves, but should more closely resemble them in the way in which they change in response to market or credit risk shocks.

In particular, I would identify two main areas in which rule waivers or modifications might be used.

  • First, we would consider waiving the "net premium" rule. Waiver of the "net premium" rule would merely permit rather than require a life insurer to include part or all of the excess of the gross over net premium in the calculation of mathematical reserves. Total inclusion of the whole gross premium in the calculation of mathematical reserves might overstate distributable surplus and could lead to adverse tax consequences. The actuary carrying out the calculation will need to exercise judgement to ensure this does not occur. One approach might be to include some provision for future annual bonuses or current surrender values in the calculation of mathematical reserves, perhaps guided by an aggregate comparison of mathematical reserves to the realistic liability. Alternatively the actuary might, in the normal way, generate the appropriate amount of surplus for distribution by transferring amounts to or from the investment reserve.
  • Second is the difficult area of the valuation of guarantees. We see two aspects that could be reconsidered. The first is our guidance on annuity guarantees, which has been criticised for setting an unreal standard. We would consider waivers here where accompanied by a more scientific approach to the time value of such options. Similarly our rules require a strict approach to persistency of policies to eventually reach and claim against the guarantees at particular dates, particularly on unitised with-profits business. This may also be an area for waivers, though a large degree of prudence is still required for guarantees that bite long into the future.

This brings me to the end of my brief outline of our proposed approach to realistic liabilities. I shall now briefly explain how this initiative relates to our other initiatives on with-profits governance and the role of actuaries. My starting point is to emphasise that due to recent developments the initiative on realistic liabilities is immediate, urgent and important. It is also related to, and impacts upon, our initiatives on governance and the role of actuaries. In particular there are some aspects of governance proposals that now become more important and other aspects of our proposals on the role of actuaries that arguably may need to be deferred. I would emphasise in particular three areas.

  • Firstly the need increases for early progress on defining the Principles and Practices of Financial Management ("PPFM") and governance arrangements, such as with-profits committees, to ensure that life insurers manage their business in accordance with their PPFM. A key aspect of the calculation of realistic liabilities is the smoothing policy. I have already mentioned that the calculation of realistic liabilities must be consistent with a life insurer’s actual practice in setting bonus levels and with how that practice has been described to policyholders. The PPFM and related governance procedures, when put in place, will play a key role in ensuring this is the case. We would encourage life insurers – to put in place a PPFM and appropriate related governance procedures as quickly as possible even in advance of our rule changes in this area. We will be looking especially to life insurers that are seeking realistic rule waivers or modifications to make early progress in this area although this will not be a formal condition for the granting of a waiver or a rule modification.
  • Secondly the need to redefine the relationship between senior management and the life actuary increases in importance. In sharp contrast to the existing regulatory rules, the calculation of the realistic financial position needs to reflect how the business is actually run. The calculation of the realistic safety margin is particularly demanding as it needs to reflect not only how the business is run in normal circumstances but also how it would be run in stressed circumstances. I have already mentioned that the bonus smoothing policy used in the calculation of realistic liabilities (under both normal and stressed circumstances) needs to be understood and approved not only by the actuaries who carry out the calculation but also by senior management. For some life insurers this means that the relationship between senior management and the life actuaries needs to change in important respects. This is another area in which we would urge life insurers to make urgent progress even in advance of our rule changes. Senior management needs to understand and take unambiguous responsibility for the work of their life actuaries. When an application for a rule waiver or modification is made we will look to senior management as well as the life actuary to make the case. When, in early 2004, we amend the regulatory rules to dispense with the need for bespoke waivers we will consider looking to the directors to certify the calculation of the realistic position. However, at this crucial time when the basis of the calculation of liabilities is undergoing material change we also believe that it will be important to pay extra care to ensure that the life actuaries remain closely involved. We are therefore minded also to keep, at least pro tem, the appointed actuary’s certificate and to extend it to cover the reporting of the realistic position. We doubt that it would be appropriate to change the source from which we gain independent assurance as to the reliability of the calculation of a life insurer’s financial position in the same year in which we change the fundamental basis of that calculation. We are minded therefore to defer this aspect of our proposed reforms.
  • Thirdly, in light of our decision at the beginning of this month to accelerate the introduction of a realistic basis to measuring financial strength, we need to be clear that our delineation of actuarial responsibilities between the actual function holder and the with-profits actuary remains appropriate.
  • The calculation of the realistic position and especially of the realistic financial position now, in effect, link the main tasks envisaged in CP167 for these two functions – advice to the Board on regulatory calculations and advice on the use of discretions – much more closely than was previously the case. Further work would may also be needed to ensure that management could not themselves sign off on the regulatory calculations contrary to the advice of the life actuary without stating in their certificate that they have departed from that advice and the reasons why they have done so. The on-going and proactive role of the actuary in identifying threats to the realistic and regulatory financial strength may also need to be re-emphasised especially in view of the relevance of this to the need to keep the amount of the realistic safety margin under regular review. Also we may need to look again at the scope of the audit (which may need to be extended even further than our CP 167 proposals) but before doing this we will need to allow time for the auditing profession to develop auditing tools relevant to the new realistic calculations.

The changes we are proposing to make to the Prudential capital system for life insurers is both significant and urgent. The procyclical dangers of the present system have been evident before our eyes and, as I highlighted in the Tiner project report last October, we need to move to a regime that is more responsive to all market conditions, recognises the elasticity in the relationship between the value of assets and the determination of liabilities and is more risk-sensitive, including operational risks. It is vital that we all – firms, the profession and the regulators - get this right and this year, 2003, will be the most critical period for laying the foundations for a strong, robust, risk sensitive system. From where I sit, I want actuaries to be a core component of multi-disciplinary teams to be focussed on this. I make that commitment from the point of view of the regulator, where I have some influence over how resources are deployed. I think, therefore, that the weight of the profession should be directed to this issue, which suggests we should defer the change to the role of the life actuary until the IPSB comes into force in 2004.

I would emphasise that these three observations are at this stage still tentative. The period for comments on our consultation paper (CP 167) does not close until 25 April. We shall only make firm decisions on the way forward after we have received those comments. In closing I would re-emphasise the importance of this consultation paper and our receiving comments from a wide range of views within the actuarial profession, across the life insurance industry more broadly and from groups which represent consumers or consumer interests.

More Speeches: