An International Comparison Of Life Assurance Solvency Standard

An international comparison of life assurance solvency standards

10 November 2005

David Hare Chris Hancorn Philippe Guijarro

DRAFT

Index

1

INTRODUCTION

4

2

ACKNOWLEDGMENTS

5

3

BACKGROUND

6

4

OUR RESEARCH

7

4.1 Selected product types 4.1.1 Life insurance protection 4.1.2 Single Premium Fixed Annuity 4.1.3 Investment contract (regular premium) with investment guarantee

9 9 9 9

4.2

Selected countries

9

4.3

Calculation principles

9

5

RESULTS AND COMMENTARY

10

5.1

Overview of Countries reserving framework

10

5.2

Life insurance protection

10

5.3

Single Premium Fixed Annuity

14

5.4

Investment contract (regular premium) with investment guarantee

17

6

INTERPRETATION / CONCLUSION

18

APPENDIX A – FULL PRODUCT TERMS

19

APPENDIX B - ASSUMPTIONS 6.1.1 Life insurance protection 6.1.2 Single Premium Fixed Annuity 6.1.3 Investment contract (regular premium) with investment guarantee

20 20 20 20

APPENDIX C - DETAILED METHODOLOGY AND APPROACH

21

APPENDIX D - COMPARISON OF SOLVENCY REQUIREMENTS BY COUNTRY

22

2

DRAFT

3

DRAFT

1 Introduction The purpose of this paper is to compare and contrast the impact of existing regulations in place for the calculation of technical provisions for life insurance business across a number of countries around the world. The authors recognise that technical provisions are only one component in the calculation of life insurers’ capital needs. However, it is hoped that the results of this study will help inform the current discussions around the proposed EU Solvency 2 Directive, as well as aid international understanding and comparability of this component of life insurer financial strength. Message to Scientific Committee: We have completed a significant amount of the work required for this paper, however certain aspects of our work continue. We expect to complete the final version of our report later in 2005 or, more likely, early in 2006, which will enable us to complete the sections highlighted as draft throughout this paper. This includes appendices, which have been excluded from this paper in order to send you something meantime to indicate the nature of our work. Whilst the sections highlighted as draft in this paper and appendices will contain more information, we have designed this draft to enable the Paris 2006 ICA reviewing committee to understand the nature of the presentation we would intend to give in Paris next year. The authors have very much enjoyed working on this project and are very excited by the prospect of presenting our findings to an international audience.

4

DRAFT

2 Acknowledgments In the course of our work we have received support from several people across a range of organisations. Without this help much of this paper could not have been produced. We are very grateful for the support we have received from: Bill Robertson Emile Smits Marijn Meijberg Steve Mills Paul Mylet In addition we would make special mention to Angus Brown, a post-graduate student from Cambridge University who provided significant support to our work and produced most of the financial results and charts provided within this report.

5

DRAFT

3 Background The regulation of life insurance companies has been undergoing a significant amount of change in a number of countries over the past decade. As might be expected, the rate and extent of change varies from country to country and can depend on a number of factors, including the existing local regulatory regime, its perceived strengths and weaknesses, or the need to improve standards in response to unexpected issues emerging for local life insurance companies. As a result of these changes, applied to past regulations which themselves had evolved over time in each individual country, an external observer might expect the impact of regulations to give rise to materially different levels of technical provisions (or reserves) being held for similar types of business written in different countries. This paper seeks to understand the extent of any actual differences by comparing the actual technical provisions required for a range of relatively simple and common life insurance products written around the world. We hope our research will provide a timely snap-shot of the impact that divergence in reserving requirements already has around the world, and the implications that this may have for the global insurance industry and local insurance markets. This is an issue which is already being considered a great deal within some groups of countries. Within the European Union, the Solvency 2 project has been created to encourage greater consistency of reserving requirements across the European Economic Community. This is being achieved by reviewing the minimum reserving and capital requirements for insurance business in a way that aims to make a sensible and consistent allowance for the risk exposures inherent in range of different insurance contracts. Whilst considerable work is continuing in the Solvency 2 project, much of this, quite rightly is looking to assess how future minimum technical provision and capital standards might be set. Rather than risk repeating work which this project will assess, we have focused on the impact of existing divergence in practice which the current Solvency 1 standards permit within Europe. We also hope our work will be of interest from a non-European perspective, as our research has extended to a number of other wider countries, including Japan, Canada and the United States of America.

6

DRAFT

4 Our research The aim of our research was to compare the degrees of prudence inherent (at individual policy level) in a number of different local solvency regulation regimes. In the first stage, policy-level calculations (i.e. technical provisions with and without any associated solvency margin that might be required) were carried out at a range of policy durations for each of three basic product types: •

Protection;

•

Annuities: and

•

Regular premium savings plans with guarantees.

In all cases, the same economic conditions were assumed for the purposes of deriving valuation assumptions, thus ensuring consistency of comparison. The results were then compared (for each product, with and without required minimum solvency margin) in order to see what conclusions could be drawn about the relative strength of different countries’ approaches. As can be seen from the graphs in section 5 of the paper, the reserve for the same policy at the same duration can vary considerably between different local solvency regimes. In order to test whether the findings were unique to the starting conditions chosen, we re-ran the comparisons for two other sets of starting conditions (one with higher yields and one with lower). These extra investigations also allowed us to compare the extent to which technical provisions in the different countries were sensitive to different economic conditions. The second stage of our research involved stochastically modelling the size of technical provision needed (at the same range of policy durations for each of the three products) in order to give a specified degree of confidence that the associated liabilities (including expenses) could be met. These monte carlo simulations were based on starting conditions equivalent to those used in the first stage. The results from the two stages were therefore directly comparable and, when combined together, quantify the extent of prudence built into the minimum policy reserves stipulated in the countries studied. At the time of writing this paper, one of the most-talked about issues to do with the EU Solvency 2 project is the “run-off” confidence level which should be aimed for within technical provisions. Figures of 50%, 60%, 75% and even 90% get discussed in various quarters. In order to help illuminate this discussion, these were the specified degrees of prudence we chose to target in our work. The detailed calculations of technical provisions, minimum solvency margins and monte-carlo simulations are covered in Appendix C.

4.1

Selected product types The products we have considered are representative of the types of business sold in many parts of the world. While we have only modelled relatively simple examples of each type, we believe the results are sufficient to give meaningful results.

4.1.1

Life insurance protection A policy under which the sum assured is paid immediately as a lump sum amount on death of the insured life during the term of the policy. The sum assured is assumed to be constant throughout the term of the policy. Fixed regular premiums are set at the outset of the policy and are paid at the start of each policy year. Regular premiums must continue in order for the

7

DRAFT

life cover to be provided for the full policy term. Failure to continue premiums results in cancellation of the policy with no surrender payment made to the policyholder. 4.1.2

Single Premium Fixed Annuity A policy under which the policyholder makes a single fixed initial investment. In return for the single initial investment, a regular fixed annual payment is made for the remainder of the policyholder’s life. The policy cannot be surrendered.

4.1.3

Investment contract (regular premium) with investment guarantee A fixed regular premium is invested in a unit-linked fund which itself invests in a mixture of equities and fixed interest securities. The unit-linked fund accumulates over the duration of the policy. At the selected maturity date of the policy, the policyholder receives a lump sum payment equal to the higher of: 1. The value of the unit-linked holding at the maturity date 2. The amount of premiums paid, accumulated at 3% compound p.a. over the life of the policy. The policy can be surrendered before maturity, in which case a surrender payment is made equal to the value of unit-linked holding at the date of surrender. While this policy is described as unit-linked, it has a number of similarities to the type of participating (or with-profit) policy that is quite common in the UK. A more detailed description of the contract terms for each of the above products is provided in Appendix A.

4.2

Selected countries In selecting the countries to be included in this study, we have aimed to give a good coverage of some of the larger EEC countries to reflect the timing of our work in relation to the Solvency 2 project. In addition, we have also included a number of non-European countries in order to give a sufficiently international perspective. The countries we have included in our work are: Canada France Germany Holland Japan Russia Spain United Kingdom.

8

DRAFT

4.3

Calculation principles In calculating the policy provisions and capital requirements, we have sought to be as consistent as possible between countries and as representative as we can of the practice in each of the countries considered. In order to achieve this, we have followed the following approaches: 1. Whilst minimum standards do exist in calculating solvency requirements for each of our selected countries, the standards are not always prescriptive for all aspects of the solvency calculations. As a result, in some cases there can be a degree of subjectivity in calculating the actual minimum technical provisions. Because of this subjectivity, solvency calculations are sometimes seen by some as an “art” (rather than a “science”), requiring skilful interpretation and application of local solvency regulations. To reflect this, we have taken support from solvency practitioners within each of our selected countries in order to understand how the current local minimum solvency standards are interpreted and applied in practice. Our discussions with the local solvency practitioners focused on understanding the methodology and approach to calculating the technical provisions as these are fundamental areas which will impact the actual technical provision calculation. 2. The economic conditions within each country are likely to be different on any valuation date. Typically this would result in a different set of future economic assumptions being used to value each contract. To avoid comparison distortions because of this, we have assumed the same economic conditions applied in all countries at 31 December 2004. The economic valuation assumptions for the calculation of the technical provisions within each country were then derived from this common source, but including whatever margin for adverse deviation was appropriate in each jurisdiction, as we outline below in 4. To reflect the impact that different underlying economic assumptions may have in a comparison of different solvency standards across our selected countries, we have repeated our comparison exercise across a range of alternative assumed economic conditions for all countries. All economic assumptions are shown in Appendix B 3. Non-economic assumptions also have a direct impact on the technical provisions. Our approach to setting non-economic assumptions has been to start with a central set of best estimate assumptions which is used for all countries. We have then included appropriate margins for adverse deviations which we cover below. 4. Whilst the approach to calculating technical provisions varies between countries, many countries work on the basis of setting best estimate assumptions, with a prudential margin provided through the application of a margin for adverse deviation. In some cases the minimum margin for adverse deviation is prescribed by regulations. In other cases, the minimum margin for adverse deviation may be determined by best practice within the local insurance community, but not formally recognised within regulations. In calculating the technical provisions, we have taken support from local solvency practitioners to understand the minimum margin for adverse deviation that might be expected by the local regulator. While we believe the allowances we have made are appropriate, it should be noted that no regulators have been consulted by us in setting them. All assumptions and margins for adverse deviation are covered in Appendix B.

9

DRAFT

5 Results and commentary 5.1

Overview of Countries reserving framework We summarise below the high level approaches to solvency calculations across our selected countries. Country

Canada France Germany Holland Japan Russia Spain United Kingdom.

Prospective of retrospective

Grouping permitted?

Eliminate future valuation strain?

Eliminate negative reserves?

Minimum Solvency margin

Assumptions prescribed or set internally?

4% reserves, 0.3% sum assured

Internally

DRAFT Prospective

Yes (provided prudent)

Yes

Yes

The information above provides a high level summary. More detail can be found in ‘Appendix D - Comparison of solvency requirements by country’.

5.2

Life insurance protection Chart 1 Comparison of technical provisions and solvency margins across a range of countries

£3,000.00 £2,500.00 £2,000.00

UK Reserve (exc solvency) UK Reserve (inc solvency)

£1,500.00

German Reserve (exc solvency) German Reserve (inc solvency)

£1,000.00 £500.00 £0.00 0

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20

Policy year

Information for other countries will be included shortly.

10

DRAFT

Chart 2a Comparison of technical provisions and solvency margins across a range of countries (low valuation interest rate) £3,500 £3,000 £2,500 £2,000

UK res 2% UK res 2% + solv margin

£1,500 £1,000 £500 £0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17 18

19

20

Policy year

Information for other countries will be included shortly.

Chart 2b Comparison of technical provisions and solvency margins across a range of countries (high valuation interest rate) £2,500

£2,000

£1,500 UK res 10% UK res 10% + solv margin £1,000

£500

£0 0

1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 Policy year

Information for other countries will be included shortly.

11

DRAFT

Chart 3 Comparison of technical reserves and solvency margins with confidence levels £3,000 99.90%

£2,500

99.50% 99.00%

£2,000

95.00%

Start Fund

90.00%

£1,500 75.00% 50.00%

£1,000 UK Reserve (exc solvency) UK Reserve (inc solvency)

£500 German Reserve (exc solvency) German Reserve (inc solvency)

£0 0

2

4

6

8

10

12

14

16

18

20

-£500 Year

Information for other countries will be included shortly.

The next chart compares the actual reserves which would be held on a German reserving basis (both with and without solvency margin) with the amount of money required to be held to cover the future outgo under the policy. This has been assessed at various confidence levels. As can be seen below, when allowance is made for the solvency margin, the total amount of capital and reserve set aside is well in excess of the 99.9% level. This is an extremely prudent overall reserving and capital requirement. Start Fund Required at Different Times (Protection, 10Yr Bonds) £1,200

£1,000

£800

99.90%

Start Fund

99.50% 99.00%

£600

95.00% 90.00% 75.00%

£400

50.00% Res erv e (ex c s olv enc y )

£200

Res erv e (inc s olv enc y )

£0 0

2

4

6

8

10

12

14

16

18

20

-£200 Year

The pattern of the reserve calculations is distorted by the requirement to eliminate negative reserves using German valuation methodology. The distorting effect occurs early in the policy life, when reserves would otherwise have been negative, and reflects the initial costs associated with the inception of the contract. The impact of the solvency margin is broadly linear, reflecting the nature of the current German solvency margin calculation.

12

DRAFT

In addition to the above calculations, we have also assessed the impact of investing in longer term bonds for assets backing the reserves over the full life of the sample policy. Whilst protection business is generally considered to contain low investment risk, it is clear from the results below that the risk is present and does reduce the level of confidence that the German reserving basis provides. As an example, the reserve (with no solvency margin allowance) broadly equates to a 50% (or best estimate) calculation. As we have not stressed expense or claim risk, the result below implies that the investment risk is large enough to absorb all the prudential margins inherent within the German reserving calculation. It is worth noting that the solvency margin is very significant, and has the effect of increasing confidence levels from approximately 50% to well over 99.9%. Start Fund Required at Different Times (Protection, 20 Year Bonds) £1,200 £1,000 99.90%

£800

Start Fund

99.50% 99.00%

£600

95.00% 90.00% 75.00%

£400

50.00% Reserve (exc solvency)

£200

Reserve (inc solvency)

£0 0

2

4

6

8

10

12

14

16

18

20

-£200 Year

13

DRAFT

Single Premium Fixed Annuity The following chart shows the evolution of reserving requirement over the duration of our sample policy on a German reserving basis. Results are shown both with and without allowance for the German solvency requirement. The results have also been reproduced to show the impact of variations in the valuation interest rate, Compared with the protection policy modelled in the previous section, the solvency margin has a relatively low impact.

Start reserve against time (German annuity, 3% escaltion, reserve + solvency margin shown by dashed line) £60,000 £50,000 2% 2%

£40,000

Reserve

5.3

5%

£30,000

5% 10%

£20,000

10%

£10,000 £0 0

10

20

30

40

50

60

Year

The next chart compares the actual reserves which would be held on a German reserving basis (both with and without solvency margin) with the amount of money required to be held to cover the future outgo under the policy. This has been assessed at various confidence levels. As can be seen below, when allowance is made for the solvency margin, the total amount of capital and reserve set aside is well in around the 95% level. So whilst this is still a relatively prudent position, it is not as prudent as that of the protection policy on a German reserving basis. The confidence levels have been calculated assuming backing assets are bonds of 10 year duration. Assets are assumed to be continually rebalanced (with no transaction costs) to maintain the 10 year duration profile throughout the run off of the business. This investment strategy provides a relatively narrow band of results across the simulations.

14

DRAFT

Fund Required at Different Times (Annuities, 10 Year Bonds) £120,000 99.90% 99.50%

£100,000

99.00% 95.00% £80,000

Start Fund

90.00% 75.00% £60,000

50.00% Res erv e (ex c s olv enc y ) Res erv e (inc s olv enc y )

£40,000

£20,000

£0 60

70

80

90

100

110

120

Age

In addition to the above calculations, we have also assessed the impact of investing in longer term bonds for assets backing the reserves over the full life of the sample policy. The valuation of annuity business is extremely sensitive to interest rates, which is demonstrated by our chart below. Our asset generator assumes that investing in longer term bonds will give rise to more volatile asset price movement over time which consequently has a significant impact on the range of results from our simulations compared with those above based on investing over a shorter duration. In addition to the increased spread of results below, it is also very interesting to observe the implied level of confidence that the German reserving basis gives. Even after allowing for the solvency margin, we are only 50% likely to have sufficient assets to cover expected future outgo when following the 30 year asset duration investment strategy. This compares with a 95% confidence level (above) when we follow a 10 year duration investment strategy. This is a clear demonstration of the lack of risk sensitivity currently within the German reserving basis.

15

DRAFT

Fund Required at Different Times (Annuities, 30 Year Bonds) £120,000

£100,000

99.90% 99.50% 99.00%

£80,000

95.00%

Start Fund

90.00% 75.00% £60,000 50.00% Res erv e (ex c s olv enc y ) Res erv e (inc s olv enc y )

£40,000

£20,000

£0 60

70

80

90

100

110

120

Age

Investment contract (regular premium) with investment guarantee The following chart shows the evolution of reserving requirement over the duration of our sample policy on a German reserving basis. Results are shown both with and without allowance for the German solvency requirement. The profile of a continually increasing reserve reflects the timing of regular premiums being paid over the life of the policy. There is also a small distortion caused by the elimination of negative reserves initially.

Reserves required against tim e (Germ an endowm ent, reserve + solvency m argin shown by dashed line)

Reserve

5.4

£16,000 £14,000 £12,000 £10,000 £8,000 £6,000 £4,000 £2,000 £0

2% 2% 4% 4% 5% 5% 0

5

10

15

20

Year

16

DRAFT

Compared with the annuity contract, the reserves are relatively insensitive to the assumed valuation interest rate. The next three charts show the amount of money required to be held to cover the future outgo under the policy under a range of alternative investment strategies. As this business is participating (or ‘with profits’) the policyholder will benefit from strong future investment returns which a more aggressive investment policy may produce. We have not considered this perspective within the results below, but are instead interested in assessing the impact of following each investment strategy on the effective solvency levels. As can be seen below, the results are extremely sensitive to the assumed investment strategy, with a 100% equity investment policy given a very wide range of results (the 99.9% point result is over 10 times that for the equivalent 10 Year Bond result below).

Insert 10 year chart Insert equities chart Insert mixed investment chart

17

DRAFT

6 Interpretation / conclusion As can be seen from the results above, there is significant variation in the level of reserves that different regulatory regimes around the world have set as their minimum requirement. Whilst this comparison has inevitably relied upon a certain element of local interpretation in assessing minimum standards, the impact of the interpretation is relatively small. Apart from the variation between countries, it has also been interesting to note the impact that alternative investment strategies can have on the likelihood of sufficiency of reserves. The results have also shown the crudeness of current solvency margin calculations, which is demonstrated by the ability to result in both extremely high and low confidence levels. The results of the analysis highlight acknowledged weaknesses of the current regulatory requirements for many companies, and the lack of comparability of solvency requirements. It also gives a good indication of the motivation that cross-border insurance/reinsurance arrangements have presented to organisations in the past. From a European perspective, Solvency 2 is seeking to reduce and/or remove the impact of divergence between member states. Whilst harmonisation of minimum reserving standards is one goal, we would view the setting of a risk based solvency margin as being an even higher priority. The results of our analysis also show that the overall impact of a change in minimum solvency requirements is not clear – some companies may see an increase in solvency and others a reduction. The mix of business written, together with current local regulatory regime will be the main drivers for the overall impact.

18

DRAFT

Appendix A – Full product terms

DRAFT

19

DRAFT

Appendix B – Assumptions

DRAFT

20

DRAFT

Appendix C - Detailed methodology and approach

DRAFT

21

DRAFT

Appendix D - Comparison of solvency requirements by country

DRAFT

22