UK: Solvency II – Implementing Measures Shedding Light On The Future Requirements

Introduction

On 2 July 2009, the Committee of European Insurance and Occupational Pensions Supervisors (CEIOPS) issued its second set of draft technical advice on Level 2 implementation measures for consultation.

Following on from the first set of consultation papers ("CPs") released earlier in March this year, this draft advice covers a range of topics across the entire Solvency II spectrum, addressing aspects of the Minimum Capital Requirement (MCR), the Solvency Capital Requirement (SCR) Standard Formula, Internal Model, Own Funds, Reporting and the Supervisory Review Process.

This draft advice will influence the final shape of insurance regulation across European Union member states, with final advice on these implementing measures due to be submitted to the European Commission in January 2010. The second set of papers is subject to a ten week consultation period ending at 4pm CET on 11 September 2009. We welcome the opportunity that this consultation process provides for the industry to engage in the development of the detailed Solvency II regulations, and we strongly encourage firms and other stakeholders to provide constructive and practical feedback.

We have presented here a summary of each of this second wave of consultation papers, pulling out the main elements of the advice, along with our view of the key messages and impacts for the UK insurance industry. We hope that you will find it useful. Deloitte has commented on all previously-issued consultation papers and we will continue to express our views to assist CEIOPS in its crucial rule-making activities.

No. 39 – Technical provisions – Actuarial and statistical methodologies to calculate the best estimate

Purpose – This paper provides detail on actuarial and statistical methodologies for the calculation of the best estimate component of technical provisions.

Key messages

For life business the key difference from current UK regulations is that negative reserves are allowed at policy level without the restrictions imposed by existing rules. We could therefore see a release of reserves, particularly for unit-linked and protection products. This, in turn, could create balance sheet volatility and a greater exposure to lapse risk.

For non-life business, separate provisions are needed for claims outstanding and for claims events occurring after the valuation date during the remaining in force period of policies held by the firm. In both cases, amounts are expected to be best estimates and to allow for the time value of money using the risk free rate determined as per CP40. In addition to the best estimate, the framework Directive requires that all technical provisions include a risk margin.

Detailed summary

Valuation process

• The valuation process should be supervised by an expert with sufficient knowledge and all steps of the process should be well documented.

Cash flow projections

• All future cash flows that would be incurred in meeting policyholders' liabilities should be considered in the calculation of the best estimate of technical provisions.
• The best estimate should be calculated gross of reinsurance and expected recoveries from reinsurers and special purpose vehicles should be shown separately on the asset side of the undertaking and should make proper allowance for expected losses due to counterparty default, whether due to insolvency, dispute or other reason.
• As far as expenses assumptions are concerned, proper allowance for expense inflation should be considered. Expense assumptions however should not allow for future cost reductions until these have been realised. The valuation should assume that companies continue to write new business.
• Due consideration of options and guarantees, policyholders' behaviour, management actions and future distribution of extra benefits should be made.

Life contracts

• A policy by policy valuation is preferred. Grouping of model points is acceptable provided that the undertaking can demonstrate that the grouping does not misrepresent the underlying risk.
• Negative best estimate technical provisions are acceptable at individual policy level, but as long as they are legally enforceable (see CP30)
• No surrender value floor should be assumed.

Non-life contracts

• Separate amounts need to be determined for outstanding claims and premium provisions.
• Premium provisions should take into account:
• • Future premium payments, taking account of future policyholder behaviour (such as policy lapse).
  • Cash flows from future claims events, including e.g. recoveries from salvage and subrogation.
  • Cash flows from allocated and unallocated claims management expenses.
  • Cash flows arising from the ongoing administration of the in-force policies.
  • Future premiums need to be taken into account (see CP30).

• Provisions for claims outstanding should include all future claims payments as well as allocated and unallocated claims management expenses whether or not claims have been notified.
• Where non-life policies give rise to payments of annuities (eg under periodic payment arrangements), claims outstanding may need to be determined using appropriate life actuarial techniques.

Options and guarantees and policyholders' actions

• Both intrinsic and time value of options and guarantees should be considered.
• The uncertainty of the cash flows should be captured taking into account multiple scenarios.
• When valuing options and guarantees, policyholders' behaviour should not be assumed to be independent of market conditions.
• Where material, non-financial guarantees should also be considered.

Management actions and valuation of discretionary benefits

• Management actions may be reflected in the best estimate.
• When valuing future discretionary benefits, future asset returns should not exceed the level given by the forward rates derived from the risk-free interest rates.

Assumptions

• Assumptions should be consistent with information provided by the financial markets and generally available data on insurance and reinsurance risks. In addition, assumptions should be derived consistently year to year.
• The asset model, which produces projections of market parameters, should be arbitrage free and should be able to replicate observable asset prices. The calibration should reflect the nature and term of the liabilities. In addition, the asset model should be calibrated to the current risk-free term structure. This paper re-opens the issue of volatility calibration in scenario generators, expressing a continued preference for implied volatilities where these can be observed.

Validation

• Appropriate validation methods should be used to ensure that the methods and assumptions used in the calculation of the best estimate are appropriate.
• The validation processes should include adequate documentation and evidence of independent review, either by internal or external experts.
• Examples of validation processes include back testing and Actual v Expected analyses.
• CEIOPS refers to stress and scenario testing in this area and incorporates the concept of reverse stress testing.

No. 40 – Technical provisions – Risk-free interest rate term structure

Purpose – This paper provides advice on the relevant risk free interest rate structure to be used in the assessment of technical provisions.

Key messages

In this paper, CEIOPS favours the use of the government bond over that of a swap yield curve for the valuation of technical provisions. The Directive requires the use of risk free rates and swaps rates make allowance for credit risk. Furthermore, CEIOPS clearly states that the majority of its members consider it inappropriate to make an allowance for illiquidity premium in the valuation of the technical provisions.

This is clearly a controversial position given the impact on, particularly, life products such as annuities. We believe that CEIOPS's position is unlikely to change even though the industry is inexhaustibly lobbying for the allowance of an illiquidity premium.

As part of this CP, the FSA presents an analysis showing that for the UK, neither the swap nor the gilt curve meets all the requirements set by CEIOPS and supports the use of swap rates less an adjustment for credit risk.

Detailed summary

The risk-free interest rate term structure must include no credit risk and it must be possible to earn these rates in practice. The rates must also be derived from a reliable and robust source and from liquid and transparent markets. Finally, the rates should not include technical bias and they should be available for all relevant currencies. CEIOPS is of the opinion that the risk free interest rate term structure as well as the methodology to derive it should be provided at least quarterly and more frequently in volatile conditions.

In practice there are four possible options for satisfying the risk-free rate criteria:

1. Government bond rates.
2. Government bond rates plus an adjustment.
3. Swap rates.
4. Swap rates minus an adjustment.

The choice of option will depend on the currency that the risk-free interest rate term structure is based on. For example, if Euro, then the government yield curve based on AAA rated government bonds and published daily by the European Central Bank should be used as this satisfies the risk-free rate criteria set out below.

UK view

In this CP, the FSA presents an analysis of the advantages and disadvantages of using the swap and gilt curves.

Technical bias

• For some years, the long term yields were lower than short-term yields due to a mismatch between supply and demand for long-dated bonds. This led to an inverted yield curve. Although the swap curve was also inverted for a number of years it is less likely to suffer from distortions as supply and demand are more linked than for gilts.
• The additional demand for "benchmark bonds" (e.g. 10 and 15 year gilts) pushes yields down and causes a lack of smoothness in the government bond curve.
• Short/medium term gilts are very liquid investments and it is possible to earn an additional risk-free return through the repo market (which is not taken into account in the publicly available gilt yield curves).
• The recent "flight to quality" creates additional volatility in the short end of the curve.
• The swap curve is less affected by monetary policy.

Credit risk-free

• Although it is generally accepted that gilts are risk-free, swap yields contain a very small margin for credit risk.

Realism

• An insurer can earn swap rates but could potentially earn more than gilts in a risk free fashion by entering a repo transaction.

Reliability

• Although gilt yields are publicly available, zero coupon swap rates will vary between investment banks. Although larger insurers could potentially receive multiple quotes and use an "average" yield curve, smaller insurers may not be able to receive any quotes at all.

Liquidity

• The swap market has traditionally been deeper and more liquid than the gilts market. However, the swap market became very illiquid during the current financial crisis.

As shown above, neither the swap curve nor the gilt curve meets all the requirements. The FSA is of the opinion that the best approach would be to use the swap rates less an adjustment for credit risk. The method to derive this adjustment is still to be decided. A practical method may be to take an average of the swap and gilt rates.

No. 41 – Circumstances in which technical provisions shall be calculated as a whole

Purpose – This paper provides advice on the circumstances under which technical provisions can be accounted for in a single calculation, rather than the sum of a best estimate provision and a risk margin.

Key messages

We believe that the definition of hedgeable risks in this consultation paper – i.e. cash flows should be perfectly replicated with financial instruments traded in a highly liquid market – will mean that most contracts will be either treated as non-hedgeable or unbundling techniques will need to be applied to separate hedgeable and non-hedgeable components of a contract if a risk margin is to be avoided for those components that are hedgeable.

Detailed summary

Technical provisions can be accounted for in a single calculation if:

• The financial instruments used to replicate the insurance and reinsurance risk provide the same uncertainty in amount and timing of cashflows in all possible scenarios.
• The market value of these financial instruments are readily observable in a highly liquid market, where these characteristics are expected to be permanent.
• Where a life contract contains several sets of cashflows, some of which do not satisfy these criteria, the technique of "unbundling" can be used. In this case:
• • The cash-flows that do satisfy these criteria can be valued using the observed financial instruments.
  • The cash-flows that do not satisfy these criteria must include a risk margin.

In practice this means that term assurance contracts will not be accounted for in a single calculation since the cashflows are dependant on life expectancy and policyholder behaviour which cannot easily be replicated by a financial instrument. Pure unit linked contracts (without any additional guarantees) can be unbundled into the number of guaranteed units and the expense cashflows. The calculation of the technical provision of the former can be treated as a single calculation and the latter using the three building blocks.

No. 42 – Calculation of the risk margin

Purpose – This paper provides detail on the calculation of the risk margin, including definitions of terms, assumptions, calibration of the Cost of Capital (CoC) rate, and projection of the future SCRs.

Key messages

The approach to be used in the risk margin calculation remains broadly unchanged from that elaborated for QIS4. The cost of capital is still 6% and there is no allowance for diversification across lines of business in the calculation. The single exception is the inclusion of the "unavoidable market risk" component of the SCR to be used in the cost of capital calculation.

Detailed summary

Guiding principles

• In calculating the risk margin, there are 10 principles used to define the "reference undertaking assumed to take over and meet" the obligations covered by the technical provisions. This "reference undertaking" concept provides a theoretical basis in support of the measures implemented in QIS4.
• These principles are provided alongside discussion of the same, including input provided as part of QIS4 particularly from the CRO Forum and the CEA.
• Of particular interest, principle 5 says the components of the SCR to be used in the cost of capital calculations should include the following:
• • underwriting risk with respect to the existing business;
  • counterparty default risk with respect to ceded reinsurance and SPVs;
  • operational risk; and
  • unavoidable market risk.

• The SCR used for the calculation of risk margin for the reference undertaking can be based on the standard formula or an internal model.

Cost-of-Capital rate

• Quantitative studies suggest that the rate should be at least 6% and should be updated annually. The three-step process used to derive this initial rate is discussed, including an in-depth discussion in Annex 1 of the consultation paper.
• A number of key conditions should be satisfied/ reflected in the determination of the cost-of-capital rate. Some of these are as follows:
• • Shareholder return models should provide the initial input.
  • The CoC rate must be a long-term average rate, reflecting both periods of stability and stress.
  • The CoC should be independent of the solvency position of the original undertaking.

Calculation of the risk margin

• The total risk margin is the sum of the risk margins for each line of business.
• For non-life business, the risk margin is not split between the outstanding claims and premium provisions elements.
• The risk margin for a line of business is calculated as the sum of the present values of the SCR for the line of business (including the components listed above) multiplied by the CoC rate, discounted using the riskfree interest rate for each year from the valuation date until the expiry of the business.

No.43 – Technical Provisions – Standards for Data Quality

Purpose – This paper provides advice on Article 85 (f) by clarifying the data quality requirements for inputs used for calculating technical provisions.

Key messages

CEIOPS continues to enforce the importance of data quality, which in our experience is an area of concern for many of our clients. It is expecting insurers to have:

1. developed a data dictionary of sources and attributes used for estimating provisions;
2. conducted the data quality assessment on this data;
3. taken steps to remediate the identified issues and plans are in place for longer term solutions; and
4. demonstrated ongoing data quality monitoring processes being in place.

Detailed summary

CEIOPS advises on how data quality should be managed with regards to Solvency II, in that the insurer must demonstrate that they have addressed the following:

• Definition of the data: Established a data dictionary, which defines attributes, granularity and sources of data used in the technical provisions calculation. The dictionary also contains definitions for historical data used in experience analysis and any externally sourced information.
• Assessment of data quality: Verified the features of the defined data elements using appropriate quality measures, covering the following criteria:
• • Appropriate – data is suitable for its intended purpose and is relevant to the risk that is being considered.
  • Complete – data is of sufficient granularity that allows for the 'main homogeneous risk groups' being identified, including any trends of risk behaviour and historical coverage.
  • Accurate – data does not contain any errors or omissions which have material impact on the provision valuation. Accurate data is captured consistently over time and the organisation demonstrates its widespread use in its operations and decision making process.

• Resolution of the problems identified: Identified data quality issues, and taken remedial action. Plans are in place for developing longer term solutions if the remedial action was tactical.
• Monitoring of the quality of data: Regular data quality monitoring processes are in place based on objective measures but also expert judgement.

In the cases where adjustments and approximations have to be made, CEIOPS advises that insurers document the event, validate the impact and set in place a long-term solution to the issue. It suggests that failure to remediate IT systems and processes does not justify approximations and manual adjustments.

CEIOPS also stresses the need for expert judgement to validate any objective data quality measures, adjustments, approximations and the usage of historical data.

No. 44 – Technical provisions – Counterparty default adjustment to recoverables from reinsurance contracts and SPVs

Purpose – This paper provides advice on the calculation of the adjustment for counterparty default from reinsurance contracts and SPVs

Key messages

There are no significant surprises in the approach suggested in this consultation paper apart from the fact that the same maximum recovery rate of 40% is recommended by CEIOPS for both the best estimate (discussed in this CP) and under stress conditions (discussed in CP51).

Detailed summary

The adjustment for counterparty default should be calculated as a multiple of the following elements:

• Probability of default of the counterparty.
• The loss-given default (which should take into account the expected recovery rate).

If the expected recovery rate cannot be accurately estimated then no rate higher than 40% should be used.

If the loss-given default takes into account risk mitigation instruments, then their associated credit risk must also be allowed for.

Ideally, the adjustment for counterparty default should be calculated separately for each line of business and counterparty. However, if the probability of default and recovery rates of several counterparties coincide then the adjustment may be calculated together.

No. 45 – Technical provisions – Simplified methods and techniques to calculate technical provisions

Purpose – This paper provides guidance for the use of simplified methodologies for the calculation of best estimate provisions in order to ensure that actuarial and statistical methodologies are proportionate to the nature, scale and complexity of the risks.

Key messages

This paper provides high-level guidelines and principles to help identify when it might be appropriate to calculate technical provisions based on simplified methods or approximations.

The CP lays out three steps, with associated examples, in the process for evaluating the technical provisions:

1. Assessing the nature, scale and complexity of the underlying risks;
2. Checking whether the valuation methodology is proportionate to the risks assessed in step 1, having regard to the degree of model error resulting from its application; and
3. Back test and validate the assessment carried out in steps 1 and 2. The CP also discusses the concept of thresholds for determining the allowance of simplified methods as well as some specific details surrounding simplifications with respect to reinsurance recoverables.

Detailed summary

This paper is primarily focused on the role of proportionality in the valuation of technical provisions.

In assessing whether a valuation method could be considered proportionate to the underlying risks, the undertaking should have regard to three steps:

1. Assessing the nature, scale and complexity of the underlying risks.

This step is intended to provide a basis for checking the appropriateness of specific valuation methods carried out in the subsequent step and serves as a guide in identifying where simplified methods are likely to be appropriate.

2. Checking whether the valuation methodology is proportionate to the risks assessed in step 1, having regard to the degree of model error resulting from its application.

In this step, an assessment is performed as to whether a specific valuation method can be regarded as proportionate to the nature, scale and complexity of the risks analysed in step 1. Also, the degree of potential error from applying the particular method is considered.

3. Back test and validate the assessment carried out in steps 1 and 2.

Finally, this step looks backwards to assess whether best estimates calculated in previous years turn out to be appropriate in subsequent years.

Simplified methods

It is seen as generally unnecessary to include detail on specific simplified methodologies for the valuation of technical provisions within the Level 2 implementing measures. Such detailed rules should only be included where:

• The use of simplified components is expected to be widespread.
• There is a particular need for small and medium-sized entities. Some specific guidance and details are provided in respect of valuation methods for reinsurance recoverables and risk margins.

No. 46 – Own Funds – Classification and eligibility

Purpose – to provide advice on the classification of own funds and their eligibility in meeting capital requirements.

Key messages

CEIOPS advises stricter criteria for eligibility and classification of eligible own funds and stricter limits on the proportions of SCR and MCR to be covered by tier 1 capital than the minimum requirements set out in the level 1 text of the Solvency II Directive (SCR – CEIOPS 50% – Directive 33.3% – MCR – CEIOPS – 80% – Directive 50%).

CEIOPS's advice that all tier 1 capital must rank pari passu to absorb losses on a going concern and winding up is likely to mean that instruments currently classified as tier 1 instruments other than core tier 1 capital may be classified as tier 2 rather than tier 1.

CEIOPS has not yet concluded on the treatment of the unearned profit element of any residual margin or composite margin that may emerge from IFRS4 Phase II.

CEIOPS's advice on ring fenced funds is not included in this CP.

To read this document in its entirety please click here.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.