STATES OF JERSEY

JERSEY TEACHERS' SUPERANNUATION FUND (JTSF): ACTUARIAL VALUATION AS AT 31ST DECEMBER 2016

Presented to the States on 16th March 2018 by the Chief Minister

STATES GREFFE

2018  R.26

INTRODUCTORY REPORT

1. Background

1. The Jersey Teachers Superannuation Fund ("JTSF") is the pension fund for teachers in Jersey. The Fund has 2,700 scheme members, with 1,170 of these members  being  active  members  employed  by  the  States  of  Jersey  and 4 Accepted Schools. The JTSF is a statutory scheme with its governance, benefits and funding arrangements defined under the Teachers' Superannuation (Jersey) Law 1979. It is a funded scheme requiring regular actuarial valuations to assess whether Fund assets are sufficient to cover long-term liabilities.
2. Under  Article 3(12)  of  the  Teachers'  Superannuation  (Jersey)  Law  1979 ("the Law"), the Management Board of the Jersey Teachers' Superannuation Fund, the Actuary shall review the operation of the Fund in respect of –

1. each 5-year period beginning on 1st September 1986; or
2. each  period,  of  less  than  5 years,  that  the  Minister  may,  after consultation with the Management Board, determine.

3. Under Article 3(13) of the Law, on each review under paragraph (12), the Actuary shall make a report to the Minister, the Minister for Treasury and Resources and the Management Board on –

1. the financial condition of the Fund; and
2. the adequacy or otherwise of the contributions payable under this Law to support the pensions and other benefits payable under this Law.

4. Whilst the scheme orders require an actuarial valuation every 5 years, it is best practice to conduct actuarial valuations of pension schemes every 3 years. The JTSF Management Board have adopted this best practice, and on the triennial cycle  the  Scheme  Actuary  has  completed  the  actuarial  valuation  as  at 31st December 2016.

2. Actuarial Valuation Result

1. The attached actuarial valuation report outlines the results of the 2016 actuarial valuation conducted by the Scheme Actuary, together with the assumptions used.  Under  Article 3(14)  of  the  Law,  a  copy  of  every  report  under paragraph (13) shall be laid by the Minister before the States as soon as practicable after it is made.
2. The main conclusions from the valuation are that –

• There is a past service surplus of £77 million as at 31st December 2016.
• The value of anticipated future contributions is less than the value of future service benefits in respect of active members as at 31st December 2016, giving rise to a future service deficiency of £42.4 million.
• Putting  these  2 elements  together,  the  Fund's  overall  surplus  as  at 31st December 2016  is  £35.1 million,  equivalent  to  a  funding  ratio  of 106.3%.

3

These results are based on the assumption that the Pension Increase Debt, amounting to £111.9 million as at 31st December 2016, for which provision has been made in the States' balance sheet, is repaid in full by the States over an appropriate period of time.

3. Over the period since the last actuarial valuation, the funding surplus has increased  from  £7.4 million  at  31st  December  2013  to  £35.1 million  at 31st December 2016. In preparing the actuarial valuation, the Scheme Actuary has updated demographic assumptions based on experience over the preceding 3 years, and also updated the market-led financial assumptions.
4. The improvement in the funding position has been due to investment returns exceeding the long-term expectations, positive demographic experience, and salary/pension increases being lower than long-term expectation. These have more than offset changes in financial assumptions resulting from the Scheme Actuary updating the market-led financial assumptions.
5. Whilst the actuarial valuation position has improved, the report highlights that final-salary pension benefits provided to teachers are unaffordable within the contribution rates currently being paid. There is a shortfall of 3.1% of teachers' pensionable salaries – around £2 million per annum for all teachers in the scheme.

6. The underfunding of the benefit package is unsustainable in the long term and is expected to result in the current surplus being depleted over time. Since the last actuarial valuation, the shortfall in contributions has increased from 2.6% of teachers' pensionable salaries to 3.1% of teachers' pensionable earnings.

3. Dealing with the Surplus

1. The  surplus  of  £35.1 million  as  at  31st  December  2016  is  based  on  the provisions of the Scheme at that date. Whilst there is a surplus relating to the past service, the shortfall in contributions to fund future benefits will reduce this surplus over time. The scheme, like other pension schemes, is experiencing increasing longevity, which is increasing the cost of pension provision. It is expected that the surplus will gradually reduce to zero over the long term, due to the shortfall of contributions that exists within the scheme. The Scheme Actuary has estimated that based on recent levels of new joiners, the emerging

R.26/2018

strain generated in relation to new entrants over the next 10 years may exceed £20 million.

2. Where a surplus is disclosed at a valuation, Article 18(4) of the Teachers' Superannuation (Administration) (Jersey) Order 2007 requires proposals to be submitted to the States to dispose of any surplus. However, given that –

• the funding strain expected when new members join the Fund, due to contributions being lower than the cost of accrual, in particular the strain in relation to new entrants over the next 10 years, may exceed £20 million in total;
• there is currently no formal agreement between the States of Jersey and the Management Board regarding the repayment of the pension increase debt; and
• there is volatility inherent in the value of the Fund's assets relative to liabilities,

the Scheme Actuary has recommended that the surplus is retained as a buffer against future adverse experience. The Employer's Actuary has confirmed it is reasonable that the surplus is retained as a prudent reserve against future adverse experience, given the future service strain that is expected to occur.

4. Summary

4.1  The JTSF actuarial valuation shows the Fund had a surplus £35.1 million, equivalent to a funding ratio of 106.3% as at 31st December 2016. It is recommended that the surplus at 31st December 2016 is retained in the Scheme as a prudent reserve against future adverse experience.

Jersey Teachers' Superannuation Fund

Actuarial valuation at 31 December 2016

Prepared for Prepared by Date Signed

JTSF Management Board Jonathan Teasdale

23 February 2018

Jonathan Teasdale FIA

jonathan.teasdale@aon.com

Copyright © 2018 Aon Hewitt Limited. All rights reserved.

aon.com

Aon Hewitt Limited is authorised and regulated by the Financial Conduct Authority. Registered in England & Wales No. 4396810

Registered office:

The Aon Centre | The Leadenhall Building | 122 Leadenhall Street | London | EC3V 4AN  This report and any enclosures or attachments are prepared on the understanding that it is  solely for the benefit of the addressee(s). Unless we provide express prior written consent  no part of this report should be reproduced, distributed or communicated to anyone else  and, in providing this report, we do not accept or assume any responsibility for any other  purpose or to anyone other than the addressee(s) of this report.  

Executive Summary

The key conclusions from the actuarial valuation at 31 December 2016 are set out below.

There was a surplus of £35.1M based on the assumptions adopted for the valuation

We have carried out a valuation of the Jersey Teachers' Superannuation Fund (the Fund) as at 31 December 2016. The purpose of the valuation is to review the operation of the Fund since the previous valuation, and to report on the financial condition of the Fund and the adequacy or otherwise of the contributions to support the benefits of the Fund.

Following advice from ourselves, the Management Board has confirmed that the assumptions adopted to determine the funding target for the Fund should be best-estimate assumptions. Under best-estimate assumptions the future outcome is just as likely to be better or worse than assumed. The rationale for using best-estimate assumptions is discussed in Appendix 4.

The main conclusions from the valuation are that:

• There is a past service surplus of £77.5M as at 31 December 2016.
• The value of anticipated future contributions is less than the value of future service benefits in respect of active members as at 31 December 2016, giving rise to a future service deficiency of £42.4M.
• Putting these two elements together, the Fund's overall surplus as at 31 December 2016 is £35.1M, equivalent to a funding ratio of 106.3%.

A key feature underlying the valuation results is that we have assumed the pension increase debt, amounting to £111.9M as at 31 December 2016, for which provision has been made in the States balance sheet, is repaid in full by the States over an appropriate period of time. There is currently no formal agreement between the Management Board and the States of Jersey regarding the repayment of the pension increase debt.

Developments since the valuation date

Since the valuation date, our update calculations suggest that the funding position has improved. This is primarily due to investment returns since the valuation date being above those assumed in the valuation. The position will be reviewed again at the next valuation.

The surplus may be retained as a buffer against future adverse experience

Article 18(4) of the Teachers' Superannuation (Administration) (Jersey) Order 2007 requires the Chief Minister, within 3 months of this report being laid before the States, to consult with the Management Board and to submit to the States proposals for disposing of the surplus. The proposals may consist of (but are not limited to) the following:

• If the surplus appears to be of a temporary nature, a recommendation that no action should be taken;
• The retention of a surplus no larger than the Actuary advises is a prudent reserve; or
• An increase in the benefits under the Fund.

Our advice to the Management Board is that the surplus of £35.1M is no larger than we would advise is a prudent reserve against future adverse experience. In particular, this takes account of the following:

• The funding strain expected when new members join the Fund, due to contributions being lower than the cost of accrual. The strain in relation to new entrants over the next 10 years may exceed £20M in total;
• There is currently no formal agreement between the Management Board and the States of Jersey regarding the repayment of the pension increase debt; and
• The volatility inherent in the value of the Fund's assets relative to liabilities.

We therefore recommend the surplus is retained as a buffer against future adverse experience.

Jersey Teachers' Superannuation Fund

Actuarial valuation at 31 December 2016 Contents

Introduction  1 Developments since the previous valuation  3 Information used  4 Valuation approach  6 Asset data  10 Valuation results  11 Reasons for change in funding position  12 New entrant cost  13 Risks and sensitivity analysis  14 Summary and conclusions  17 Appendix 1: Scope of advice  19 Appendix 2: Provisions of Fund  20 Appendix 3: Membership data  25 Appendix 4: Rationale for best-estimate assumptions  27 Appendix 5: Valuation method  29 Appendix 6: Financial assumptions  30 Appendix 7: Demographic assumptions  34 Appendix 8: Summary of assumptions  40 Appendix 9: Discontinuance test  42 Glossary  45

Introduction

This report has been prepared for the Management Board. It considers the financial position of the Fund as at 31 December 2016.

Legislation

In accordance with Article 3(13) of the Teachers' Superannuation (Jersey) Law 1979, we have carried out a valuation of the Jersey Teachers' Superannuation Fund (the Fund) as at 31 December 2016.

Under the legislation, valuations of the Fund are required at least once every five years. However, it is the policy of the Management Board to require valuations once every three years so as to keep the finances of the Fund under more frequent scrutiny.

The results of the valuation are based on the Orders of the Fund in force at the date of signature of this report.

Purpose

The purpose of the valuation is to review the operation of the Fund since the previous valuation, and to report on the financial condition of the Fund and the adequacy or otherwise of the contributions to support the pensions and other benefits of the Fund.

Previous valuation

GAD's valuation report dated 7 January 2015 considered the financial position of the Fund as at 31 December 2013.

Contributions since the previous valuation

Since the previous valuation contributions have continued to be paid at the rates specified in the Fund's Orders.

Next valuation

In accordance with the policy of the Management Board, the next valuation is due to be carried out as at 31 December 2019.

Scope of advice

The report is prepared for the Management Board. Please see Appendix 1 for further details of the scope of advice.

Words used

Our report includes some technical pension terms. The words shown in bold print are explained further in the glossary.

For brevity, we have also used the following shorthand: Shorthand  What it means

Snapshot view

The report concentrates on the Fund's financial position at the valuation date. As time moves on, the Fund's finances will fluctuate. If you are reading this report some time after it was produced, the Fund's financial position could have changed significantly.

Developments since the previous valuation

This section summarises the key developments since the previous valuation.

The financial health of the Fund depends fundamentally on how much cash is paid in, how well the assets perform, and on what benefits are paid out. The key developments since the previous valuation therefore include:

• The amount of contributions paid to the Fund.
• The actual returns on the Fund's investments.
• Whether there are changes to future expectations of benefit payments or investment returns.

These items are discussed later in this report. As well as these high level points, please note the developments below.

Dealing with the 2013 valuation surplus

The valuation as at 31 December 2013 revealed a surplus of £7.4M based on future increases in pensions and deferred pensions in line with the annual increase in the Jersey RPI.

Where a surplus is disclosed at a valuation, the Orders governing the Fund require proposals to be submitted to the States to dispose of any surplus. It was agreed that the surplus in the Fund be retained as a buffer against future adverse experience.

Changes to Fund Orders

The New Members Orders have been amended from 23 December 2016 to allow members in this category to exchange up to 30% of their pension for a lump sum at retirement (as opposed to 25% previously).

Information used

The information used for the valuation is summarised below.

To carry out the valuation, we have obtained information on:

• The assets held by the Fund.
• How benefit entitlements are calculated.
• Member data.

This section sets out a high level summary of the information used. Further details are included in Appendices 2 and 3.

Assets

The Scheme's assets had an audited market value of £484.6M at the valuation date. For further details, please see the Asset Data section.

Benefits valued

Members are entitled to benefits defined in the Orders. We are not aware of any established practice of granting additional discretionary benefits and no allowance for such benefits has been made in this valuation. A summary of the benefits valued is set out in Appendix 2.

Pension increase debt

The Fund provides pensions and other benefits which are subject to increases based on the rate of inflation. Prior to 2007, the cost of these pension increases was met on a "pay-as- you-go" basis by the employer as opposed to being funded in advance within the Fund.

As a result of a decision by the States of Jersey effective in 2007 to pay pension increases (both those already awarded, and those due in future) from the Fund, a shortfall arose in respect of the indexation of benefits accrued before 2007. The total deficit in the Fund was assessed as £91.6M as at 31 December 2010. This deficit became known as the 'pension increase debt' and it was recognised (via a provision in the States balance sheet) that this debt should be paid into the Fund over a period of time.

The debt should vary over time to reflect interest on the debt (which would increase the debt) and additional contributions (which would reduce the debt). However, the precise details of such a mechanism have yet to be agreed between the Management Board and the States of Jersey. We have calculated the pension increase debt as at 31 December 2016 to be £111.9M based on the proposals submitted by the Management Board to the States of Jersey in 2012.

For the purpose of this valuation, we have taken account of the pension increase debt as an asset of the Fund in line with its calculated value at 31 December 2016 of £111.9M.

Whilst the mechanics of the debt repayment mechanism have not yet been formally agreed between the Management Board and the States, the States has publicly recognised a provision within its annual accounts for the debt to be settled in full and so it seems

reasonable to rely on this for the purposes of the valuation.

The Management Board needs to be aware that, if it becomes clear at a future valuation that part or all of the debt will not ultimately be repaid, this may have a significant adverse effect on the Fund's financial position and, potentially, on members' future pension increases from that point on.

Membership data

The valuation calculations use membership data supplied by the Dedicated Pensions Unit of the States Treasury Department at 31 December 2016.

The following chart illustrates how the membership profile is evolving. Please see Appendix 3 for a more comprehensive summary of the data.

Active Members Deferred Pensioners Pensioners

1091 1150 1176 1048

919

765

406 479 502

2010 2013 2016 2010 2013 2016 2010 2013 2016

The illustration shows that there has been an increase in the number of members in each of the membership categories since the last valuation.

Reliability of information

We have carried out general checks to satisfy ourselves that:

• The information used for this valuation is sensible compared with the information used for the previous valuation.
• The results of this valuation can be traced from the results of the previous valuation. However, the results in our report rely entirely on the accuracy of the information supplied.

Valuation approach

This section describes the approach taken for the valuation calculations. Adequacy of contributions

The contributions to the Fund are specified in the Orders governing the Fund and are paid so as to provide the benefits which will become payable to members when they retire or otherwise leave the Fund.

The factors affecting the Fund's finances are open to changing circumstances. Consequently it is necessary to review the operation of the Fund from time to time, by means of an actuarial valuation, to determine the adequacy or otherwise of the contributions to support the benefits payable under the Fund.

Funding target and funding objective

In our review we start with the known facts about the Fund at the valuation date, i.e. the benefit and contribution structure, the membership and the assets. We then must make assumptions about the factors affecting the Fund's future finances such as investment returns, pay increases and rates of mortality, leaving service and retirement.

In order to calculate the value placed on the benefits, the benefits paid out by the Fund are estimated for each year into the future. The estimated benefit payments are then 'discounted back' to the valuation date using an assumed investment return known as the discount rate.

The benefit payments from the Fund are expected to be made for a very long period and Fund cashflows are linked to future levels of inflation – the chart below shows the cashflow pattern for a typical pension scheme.

For the purpose of assessing whether the contributions are adequate to support the current benefits, it is appropriate to set a "funding target" and "funding objective".

The terms "surplus" and "deficiency" are referred to in the Fund's Orders but are not explicitly defined. In practice, we say there is a surplus if the funding target is more than fully met and we say there is a deficiency if the funding target is less than fully met.

The Management Board has determined the following funding target and funding objective:

• The funding target is that, based on best estimate assumptions, the assets and future contributions (including any contributions allocated to meet the pension increase debt) should be sufficient over the long term to support the benefits payable from the Fund in respect of the current members of the Fund.
• The funding objective is that the funding target should be met and that any variations in outcome should be dealt with following each valuation in accordance with the Orders, by adjustments to contributions and/or benefits or by carrying forward surpluses and deficiencies where appropriate.

Under best-estimate assumptions the future outcome is just as likely to be better or worse than assumed. The rationale for using best-estimate assumptions is discussed in Appendix 4.

For the purposes of assessing suitable assumptions at this valuation, the Management Board agreed that the Actuary should make allowance for continued future investment in growth assets, such as equities, by assuming that liabilities will be backed by assets in line with the Fund's strategic benchmark at the valuation date (as summarised within Appendix 6).

Changes from previous valuation

The funding objective is unchanged from the previous valuation although there have been changes to the assumptions used, as discussed below.

We have changed the approach to the valuation of the pension increase debt. At the previous valuation the Actuary used a discounted cashflow approach to value an assumed stream of future contributions allocated to meet the debt. At this valuation we have taken the pension increase debt into account on the valuation balance sheet equal to its outstanding value at 31 December 2016 of £111.9M, i.e. the amount recognised as a provision in the States accounts at that date.

Valuation method

A description of the method used for the valuation calculations is set out in Appendix 5.

For the previous valuation, the Management Board specified that contributions equal to 5.6% of salaries should be allocated to meet the pension increase debt, leaving employer contributions of 10.8% of salaries to cover the cost of future benefit accrual and administration expenses.

This approach has been retained for this valuation. The contributions to meet the pension increase debt are included in the value of assets used in calculating the past service surplus / deficiency, and employer contributions equal to 10.8% of salaries are allowed for in calculating the future service surplus / deficiency.

Valuation assumptions

The results of a valuation are very sensitive to the assumptions made. The financial assumptions have a significant effect on the results of a valuation. However, the other assumptions, particularly the mortality assumptions, are also very important.

Use of market-led financial assumptions

We have adopted a market-led approach, which involves:

• market-led financial assumptions for valuing the liabilities and future contributions; and
• valuing the assets at market value.

Key financial assumptions

The following table shows the key financial assumptions used for this valuation, with the assumptions used for the previous valuation shown alongside for comparison. Important points to bear in mind are:

• The differences between the rates have a bigger impact on the results of the valuation than the absolute levels of each assumption.
• The assumptions were derived from market yields at the valuation date to ensure compatibility with the market value of the assets.

Full details of the financial assumptions used for this valuation, and the reasons for the changes compared to the previous valuation, are set out in Appendix 6 to this report.

Comparison of financial assumptions with 2013 valuation

Overall (ignoring any changes to the demographic assumptions), the financial assumptions we have used result in a lower surplus than if the assumptions used for the 2013 valuation had been retained. The main reason for this is the reduction in the discount rate (relative to Jersey RPI) used to value the liabilities.

Demographic assumptions

Other important assumptions used to value the liabilities include:

• the assumed future rates of mortality;
• the allowance made for the extent to which members will choose to exchange pension for a cash lump sum at retirement (at the rate of £13.50 cash lump sum for each £1 annual pension given up);
• the allowance for additional increases to salaries due to promotion, service or seniority; and
• the allowance made for the age at which members in each membership category will retire in future.

Comparison of demographic assumptions with 2013 valuation

We have reviewed the extent to which the demographic assumptions adopted for the 2013 valuation of the Fund remain appropriate for the current valuation as at 31 December 2016 after analysing the experience of the Fund during the 3 years 2014-2016 and taking account of other relevant data. Full details of the demographic assumptions used for this valuation, and the reasons for any changes compared to the previous valuation, are set out in Appendix 7 to this report.

In the light of this review we have made a number of changes to the demographic assumptions. The overall effect of these changes is to increase the surplus. The most financially significant change we have made is to reduce the allowance for future improvements in post-retirement mortality rates.

General comments on the assumptions

In our opinion, the financial and demographic assumptions, taken as a whole, are an entirely reasonable best-estimate basis for assessing the funding position of the Fund.

Asset data

The audited accounts for the Scheme for the year ended 31 December 2016 show the assets were £484.6M.

The Fund's assets are held separately from those of the States of Jersey. The audited Fund accounts for the year ended 31 December 2016 show its assets as £484.6M.

The assets can be categorised as follows:

Absolute return bonds: £43.2M

Property: £90.9M

Hedge funds: £62.2M

Cash and net current assets: £2.0M

Equities: £286.3M

Valuation results

Based on the assumptions set out in the Valuation Approach section, the Fund surplus at 31 December 2016 is £35.1M, equivalent to a funding ratio of 106.3%.

A detailed breakdown of the results of the main valuation calculations is given below.

The above table shows there is a past service surplus of £77.5M at 31 December 2016.

There is also a future service deficiency of £42.4M, i.e. the value of future contributions anticipated from employers and employees (disregarding any component of the employer contributions assumed to be allocated to meet the pension increase debt) is £42.4M less than the value of future benefit accrual in respect of current active members.

Putting this together, the overall Fund surplus at 31 December 2016 is £35.1M, equivalent to a funding ratio of 106.3%.

Reasons for change in funding position

The funding position has improved from a surplus of £7.4M at 31 December 2013 to a surplus of £35.1M at 31 December 2016.

The chart below shows the key reasons for the change in funding position between 31 December 2013 and 31 December 2016.

£M

Surplus/ (Deficit) brought forward 7.4

Interest on surplus 1.7

Investment returns 44.6 Deferred revaluation experience 2.9

Pension increase experience 16.7 Salary increase experience 15.8

Changes to financial assumptions -70.2

Changes to demographic assumptions 21.4

Miscellaneous -5.2

Surplus/ (Deficit) carried forward 35.1

The analysis shows that the main factors which have contributed to the improvement in the funding position since the previous valuation have been:

• Higher than expected investment returns; and
• Lower actual pension and salary increases than assumed at the previous valuation.

These factors have been partially offset by the changes in financial assumptions which have worsened the position.

New entrant cost

The current contribution rate is insufficient to meet the cost of future benefits for new joiners to the Fund.

The table below shows the contribution rate required to meet the cost of future benefits for new joiners to the Fund over the year following the valuation date, compared to the contributions expected to be received (which are fixed in the Orders).

For joiners in future years, the rate required is expected to gradually increase as continued improvements in life expectancy take effect.

Based on recent levels of new joiners (the total annual salary for new joiners in 2016 was £3.4M), the emerging strain on the Fund generated in relation to new entrants over the next 10 years may exceed £20M.

The future is highly uncertain but if all our assumptions were borne out after the valuation date and there are no changes to the assumptions at future valuations, we would expect the surplus of £35.1M at the valuation date to be eliminated over time.

Assumptions

The assumptions used to calculate the cost of future benefits above are consistent with those used to calculate the main valuation results. In addition, we have assumed the following:

• New entrants are aged 32 on joining the Fund (this is an expected average age); and
• Two-thirds of the new entrants are females and one-third are males.

Risks and sensitivity analysis

The Fund faces a number of key risks which could affect its funding position.

This section comments on some of the key risks faced by the Fund. It concentrates on the deterioration to the Fund's finances that may arise in various hypothetical downside scenarios (where the actual experience is less favourable than the assumptions made at this valuation).

However, as the assumptions used to determine the funding target are best-estimate assumptions, it needs to be recognised that upside scenarios (where the experience is more favourable than the assumptions) are just as likely.

Key risks

Here is a recap of some of the key factors that could lead to deficiencies in future:

• Investment performancethe return achieved on the Fund's assets may be lower than allowed for in the valuation.
• Investment volatilitythe assets may not move in line with the value of benefits. The Fund invests in assets (e.g. equities) that are expected to achieve a greater return than the assets (i.e. index-linked gilts and investment grade derivatives) that most closely match the expected benefit payments. The less matched the investment strategy is, the greater the risk that the assets may not move in line with the value of benefits.
• Mortalitymembers could live longer than foreseen, for example, as a result of a medical breakthrough. This would mean that benefits are paid for longer than assumed, resulting in a higher cost of providing the benefits.
• Options for Membersmembers may exercise options resulting in unanticipated extra costs. For example, members could swap less of their pension for cash at retirement than is assumed.

Quantifying the risks

To help the Management Board understand the susceptibility of the funding position on the valuation assumptions, we have considered the hypothetical impact on the liabilities of the following one-off step changes.

• Life expectancy at age 60 is three years greater than anticipated (with corresponding increases at other ages).
• The market value of equities and alternatives falls by 25% (with no change in bond markets).
• Inflation is 1% p.a. higher than assumed (with nominal expected returns on the Fund's assets unchanged).

Please see the chart below for the results.

Initial funding ratio 106% What iflife expectancy increases by three years 97%

What ifequities and alternatives fall by 25% 90%

What ifinflation increases by 1% p.a. 83%

The accrued benefits position is similarly sensitive to these factors.

The scenarios considered are not "worst case" scenarios, and could occur in combination (rather than in isolation). Opposite step changes, such as what happens if inflation falls by 1% p.a. for example, would improve the funding position by broadly similar amounts to the reductions identified above.

Investment strategy

The Fund's liabilities are influenced by Jersey inflation either directly via pension increases or indirectly via pay increases. The assets that most closely match the Fund's liabilities are index-linked gilts and investment grade derivatives. However, a large proportion of the Fund's assets are invested in asset classes such as equities which are expected to produce higher returns over the long term than those more closely matching assets.

The Management Board recognises the degree of risks, as well as the potential rewards, that this holds for the Fund. In particular the financial position of the Fund can be affected by sudden (or gradual) changes in market values of return seeking assets, changes in expected future returns and/or changes in inflation.

The investment strategy of the Fund is set by the Management Board and is kept under regular review.

Pension increase debt repayments

The valuation results rely on the pension increase debt being repaid over an appropriate period.

Implications

The analysis in this section emphasises that the Fund is highly susceptible to:

• Equity markets falling or inflation expectations rising. This risk arises because the Fund is not invested in the assets that most closely match the expected future cashflows (i.e. index-linked gilts and investment grade derivatives); and
• Members living longer than expected.

Summary and conclusions

The surplus in the Fund at 31 December 2016 is £35.1M, equivalent to a funding ratio of 106.3%. We recommend that this surplus is retained as a buffer against future adverse experience.

The headlines at the valuation date are:

• There is a past service surplus of £77.5M.
• The overall surplus, after allowing for the anticipated shortfall in future contributions, is £35.1M. This corresponds to a funding ratio of 106.3%.

A key feature underlying the valuation results is that we have assumed the pension increase debt, amounting to £111.9M as at 31 December 2016, for which provision has been made in the States balance sheet, is repaid in full by the States over an appropriate period of time. There is currently no formal agreement between the Management Board and the States of Jersey regarding the repayment of the pension increase debt.

The surplus of £35.1M will need to be dealt with in accordance with the terms of the Fund's Orders.

Developments since the valuation date

Since the valuation date, our update calculations suggest that the funding position has improved. This is primarily due to due to investment returns since the valuation date being above those assumed in the valuation. The position will be reviewed again at the next valuation.

Dealing with the surplus

Article 18(4) of the Teachers' Superannuation (Administration) (Jersey) Order 2007 requires the Chief Minister, within 3 months of this report being laid before the States, to consult with the Management Board and to submit to the States proposals for disposing of the surplus. The proposals may consist of (but are not limited to) the following:

• If the surplus appears to be of a temporary nature, a recommendation that no action should be taken;
• The retention of a surplus no larger than the Actuary advises is a prudent reserve; or
• An increase in the benefits under the Fund.

Our advice to the Management Board is that the surplus of £35.1M is no larger than we would advise is a prudent reserve against future adverse experience. In particular, this takes account of the following:

• The funding strain expected when new members join the Fund, due to contributions being lower than the cost of accrual. The strain in relation to new entrants over the next 10 years may exceed £20M in total;
• There is currently no formal agreement between the Management Board and the States of Jersey regarding the repayment of the pension increase debt; and

• The volatility inherent in the value of the Fund's assets relative to liabilities.

We therefore recommend the surplus is retained as a buffer against future adverse experience.

Appendix 1: Scope of advice

This report is prepared under the terms of the Actuary Agreement dated 14 July 2016 between Aon Hewitt Limited and the Management Board, on the understanding that it is solely for the benefit of the addressee.

Unless prior written consent has been given by Aon Hewitt Limited, this report should not be disclosed to or discussed with anyone else unless they have a legal right to see it.

Notwithstanding such consent, Aon Hewitt Limited does not accept or assume any responsibility to anyone other than the addressee of this report.

Appendix 2: Provisions of Fund

Orders

The Fund is governed by Orders made under the Teachers' Superannuation (Jersey) Law 1979 (as amended). At the valuation date, the provisions of the Fund were specified in the following Orders, namely:

1. The Teachers' Superannuation (Existing Members) (Jersey) Order 1986 – known as the Existing Members Orders
2. The Teachers' Superannuation (New Members) (Jersey) Order 2007 – known as the New Members Orders
3. The Teachers' Superannuation (Administration) (Jersey) Order 2007 – known as the Administration Orders

Main features

The main features of the Fund in force at the valuation date are summarised on the following pages.

Death in Service  1.  Cash sum:  1.  Cash sum:

1. Less than 5 years'  a)  Less than 5 years' reckonable service: a cash  qualifying service: a cash sum of 2/5ths of current  sum of 2/5ths of current salary for each year of  salary for each year of service  service
2. At least 5 years' reckonable  b)  At least 5 years' qualifying service: a cash sum of twice  service: a cash sum of current salary  twice current salary

2. Spouse's Pension (widow /  2.  Spouse's Pension (widow / widower / civil partner):  widower / civil partner):

50% of member's pension  Subject to 2 years' qualifying based on the pension which the  service: 50% of member's member would have received  pension, based on salary at on immediate ill-health  death and pensionable retirement. Pensions are only  service to normal retiring age payable to widowers in respect

of service after 6 April 1988.

An increased pension is payable

for the first 3 months after the

date of death (in most cases).

3. Dependant's Pension:  3.  Dependant's Pension:

An amount equal to a spouse's  Subject to 2 years' qualifying pension may be paid to an adult  service: an amount equal to a dependant – except that no  spouse's pension may be dependant's pension can be  paid to an adult dependant – awarded where a spouse's or  except that no dependant's child's pension is payable  pension can be awarded

where a spouse's pension is payable

4. Children's Pension:  4.  Children's Pension:

A pension is payable to each  Subject to 2 years' qualifying eligible child. The total payable  service, a pension is payable is restricted to the equivalent of  to each eligible child. The the spouse's pension, but no  total payable is restricted to one child may receive more  the equivalent of the

than half that sum. The child's  spouse's pension, but no one pension is increased if a  child may receive more than spouse's pension is not payable  half of that sum. The child's

pension is doubled if a spouse's or dependant's pension is not payable.

Death after  1.  Spouse's Pension (widow /  1.  Spouse's Pension (widow / Retirement  widower / civil partner):  widower / civil partner):

50% of member's pension in  From date of death, 50% of respect of service for which a  member's pension, ignoring spouse's pension is payable  any reduction for lump sum according to Article 64. An  taken at retirement increased pension is payable for

the first 3 months after the date

of death (in most cases).

2. Dependant's Pension:  2.  Dependant's Pension:

An amount equal to a spouse's  An amount equal to a pension may be paid to an adult  spouse's pension may be dependant – except that no  paid to an adult dependant – dependant's pension can be  except that no dependant's awarded where a spouse's or  pension can be awarded child's pension is payable  where a spouse's pension is

payable

3. Children's Pension:  3.  Children's Pension:

A pension is payable to each  A pension is payable to each eligible child. The total payable  eligible child. The total

is restricted to the equivalent of  payable is restricted to the the spouse's pension, but no  equivalent of the spouse's one child may receive more  pension, but no one child than half that sum. The child's  may receive more than half pension is increased if a  that sum. The child's pension spouse's pension is not payable  is doubled if a spouse's or

dependant's pension is not payable

4. Cash sum  4.  Cash sum For members with less than 10  None

years' reckonable service, a

lump sum equal to 5 times the

annual pension less the sum of

the pension payments made to

the member. For members with

10 or more years' reckonable

service, a lump sum equal to the

member's average salary less

the sum of the pension and

cash payments made to the

member.

Appendix 3: Membership data

Active members at 31 December 2016 (31 December 2013)

Notes:

1. The average ages shown above are unweighted.
2. Figures shown are full-time equivalent salaries at 31 December.
3. Average service includes added years arising from additional voluntary contributions.

Deferred pensioners at 31 December 2016 (31 December 2013)

Notes:

1. The average ages shown above are unweighted.
2. The 2016 pension amounts shown above include the 1 January 2017 pension increase, but the 2013 figures do not include the 1 January 2014 pension increase.

Pensioners at 31 December 2016 (31 December 2013)

Notes:

1. The average ages shown above are unweighted.
2. The 2016 pension amounts shown above include the 1 January 2017 pension increase, but the 2013 figures do not include the 1 January 2014 pension increase.
3. "Dependants" consists of spouses, civil partners, children and adult dependants in receipt of a pension.

.

Appendix 4: Rationale for best-estimate assumptions

Best-estimate assumptions

Following advice from ourselves, the Management Board has confirmed that the assumptions adopted to determine the funding target should be best-estimate assumptions. The rationale for using best-estimate assumptions is discussed below.

Range of assumptions

The results of a valuation are sensitive to the assumptions made and therefore the choice of appropriate assumptions is important.

There is a wide range of assumptions that could be used ranging from optimistic, through best-estimate to cautious:

• Under optimistic assumptions the future outcome is more likely to be worse than assumed;
• Under cautious assumptions the future outcome is more likely to be better than assumed;
• Under best-estimate assumptions the future outcome is just as likely to be better or worse than assumed.

The Management Board has a duty to protect members' benefits. Therefore it would not be appropriate to use optimistic assumptions when determining the adequacy or otherwise of the contributions to support the benefits payable under the Fund.

This leaves a choice of assumptions in the range from best estimate to cautious. The more cautious the valuation assumptions, the greater the valuation liabilities will be and consequently the greater the possibility of members' benefits or future pension increases having to be cut back (or members' or employers' contributions having to be increased) if there is a deficiency.

Advantages of using best-estimate assumptions (and disadvantages of using more cautious assumptions)

The advantage of using best-estimate assumptions is that it complies with the principle of only cutting back on the members' pensions where this appears genuinely necessary.

Using more cautious assumptions would lead to a larger deficiency, which may potentially trigger reductions to benefits or future pension increases (or increases to members' or employers' contributions). In the long term, given the extra returns targeted under the Fund's investment strategy, there would be quite a high probability that experience would prove more favourable than assumed, leading to surpluses at later valuations. Therefore, using more cautious assumptions may result in cutting back benefits (or increasing contributions) in a way that with hindsight was unnecessary.

Disadvantages of using best-estimate assumptions (and advantages of using more cautious assumptions)

The disadvantage of using best-estimate assumptions is that it leads to a larger chance of actual Fund experience being worse than assumed than if more cautious assumptions are used. This increases the likelihood of deficiencies arising at later valuations which have to be dealt with through future reductions in benefits, or by increasing members' or employers' contributions. If experience is adverse, the reductions in benefits (or increases in contribution) eventually required may need to be bigger at that time than if they had been made earlier (and therefore impacting disproportionately on a later "generation" of members). Although there is no provision in the Orders for the Fund to be discontinued, this could be particularly problematic if the Fund were discontinued. It could be equally problematic if the financial strength of the States of Jersey were to become poor. Significant benefit reductions may be required in such situations.

A further potential disadvantage of using best-estimate assumptions is that it involves anticipating a degree of outperformance from growth assets, which may limit the Management Board's scope to reduce the Fund's investment allocation to growth assets in future.

Recommendation

Following advice from ourselves, the Management Board has confirmed that the assumptions used to determine the funding target should be best-estimate because:

• It complies with the principle of only cutting back on members' pensions where this appears genuinely necessary, and
• The Management Board does not currently consider the financial strength of the States of Jersey to be poor.

Appendix 5: Valuation method

Valuation method

The valuation method for the main valuation calculations is known as the "aggregate funding" method. To establish whether the funding target is met, we have compared the value of the benefits payable in respect of all current members (including pensioners and deferred pensioners), with the sum of the following:

• the value of the Fund's existing assets;
• the pension increase debt; and
• the value of future contributions due from and in respect of current active members.

This approach involves taking credit for the future pension increase debt repayments. For the previous valuation, the Management Board specified that contributions equal to 5.6% of salaries should be allocated to meet the pension increase debt, leaving employer contributions of 10.8% of salaries to cover the cost of future benefit accrual and administration expenses.

This approach has been retained for this valuation. The contributions to meet the pension increase debt are included in the value of assets used in calculating the past service surplus / deficiency, and employer contributions equal to 10.8% of salaries are allowed for in calculating the future service surplus / deficiency.

Value of liabilities and future contributions

To calculate "the value" of the benefits payable we use our assumptions to estimate the payments which will be made from the Fund throughout the future lifetimes of current members, pensioners, deferred pensioners and their dependants. We then calculate the amount of money which, if invested now, would be sufficient to make these payments in future, using our assumptions about investment returns. The same technique is adopted to value future contributions to the Fund.

Value of assets

We have taken the assets into account at their market value.

Appendix 6: Financial assumptions

Introduction

In this appendix we describe the financial assumptions. The financial assumptions that have been chosen are consistent with the funding target and each assumption is intended to represent a reasonable best estimate of the future.

When assessing a set of financial assumptions, greater importance should be attached to the relative differences between the assumptions, rather than to the individual assumptions in isolation. This is because the differences have a greater effect on the results of the valuation than the absolute values of each assumption.

Discount rate (investment return)

The most important individual assumption in terms of its impact on the overall valuation results is the choice of discount rate, i.e. assumed future investment returns. The discount rate is used to value payments due out of the Fund (benefit payments) and into the Fund (future contributions).

For valuing the liabilities, an assumption which could be described as "low risk" would be to discount future benefit payments at the market yields available on index-linked gilts at the valuation date. This approach recognises that a good matching asset for the Fund's cash flows is obtained by investing in index-linked gilts of appropriate term.

It is common for UK occupational schemes to adopt a funding target which incorporates a higher discount rate than the returns available on gilts. The consequence of using a higher discount rate is that a lower funding target is adopted. This does not mean that the actual cost of providing the benefits is reduced, but it does result in an increase in disclosed surpluses or decrease in disclosed deficiencies.

The funding target adopted requires that the assumptions chosen should be reasonable best estimates. In principle, we need to set the discount rate at this valuation by considering the best estimate returns available on the Fund's invested assets, over the period starting now and ending in the long-term future. The expected returns depend critically on what asset classes are assumed to be held.

In line with the assumption made at the 2013 valuation, the Management Board has agreed to assume that the long-term investment strategy is the same as the strategic benchmark at the valuation date (with assets held within each class consistent with those held at the valuation date).

The best-estimate returns assumed for each asset class are set out in the table below. After applying a weighted average, this results in an assumed discount rate for the 2016 valuation of 5.7% p.a..

* Weighted average of median returns on component asset classes. The median returns on those asset classes represent best-estimate returns as at the valuation date over the 30 year period from the valuation date.

The same discount rate has been used for valuing future contributions. Increases to pensions in payment and deferred pensions

The Fund provides for annual increases to pensions in payment and deferred pensions in line with increases in the Jersey RPI.

The Bank of England produces data, based on UK fixed and index-linked gilt markets, which can be used to calculate market-implied ("break-even") UK RPI inflation. At 31 December 2016, the single break-even UK RPI inflation assumption that would give approximately the same value of liabilities as using the full Aon Hewitt UK RPI curve is 3.4% p.a..

Aon Hewitt's view is that at the valuation date, break-even inflation over the duration of the liabilities overstates likely inflation over that period, due to supply/demand distortions in the gilt market. Our best estimate is that actual inflation over the duration of the liabilities will be around 0.3% p.a. below break-even inflation (this difference is called an "inflation risk premium"). We have allowed for this in the valuation.

We have therefore assumed increases in UK RPI inflation will be 3.1% p.a..

In deciding on an appropriate assumption for Fund increases, it is necessary to take a view on the likely relationship between Jersey RPI inflation and UK RPI inflation.

The assumption for Jersey RPI for the 2016 valuation reflects the following thinking:

• The two economies have a tied currency and the same interest rates and so over the medium to long term, underlying Jersey inflation can be expected to be fairly close to UK inflation.
• The longer term past experience suggests inflation may be harder to control in Jersey. Over the period between 1949 and 1989 the average inflation rates of the two economies were almost identical. Since 1990 there have been periods where Jersey inflation has been considerably higher than UK inflation but in recent years Jersey inflation has been lower than UK inflation.

• The States of Jersey's Chief Statistician has previously indicated that he believed the calculation basis for the Jersey RPI changed materially in 1989 (Jersey RPI might have been around 0.5% p.a. lower prior to 1989 had the current approach been used).
• Due to the different calculation methodologies for calculating RPI in Jersey and in the UK, assuming Jersey RPI is equal to UK RPI is consistent with assuming that underlying inflation in Jersey will be higher than in the UK. Statistical authorities (including the Office for National Statistics who are required to publish the UK RPI) regard the present method of calculating the UK RPI as statistically flawed. The UK CPI is a preferred measure of inflation in the UK.

Given that we would expect Jersey RPI to be lower than UK RPI due to the different calculation methodologies for calculating RPI and consistent with actual experience since 2005, we have assumed for this valuation that Jersey RPI will be on average equal to UK RPI less 0.25% p.a.. (this is consistent with assuming that underlying inflation in Jersey will be slightly higher than in the UK). This differs from the previous valuation when it was assumed that Jersey RPI inflation would be on average equal to UK RPI inflation.

The assumption for Jersey RPI is therefore 2.85% p.a.. General salary increases

The States of Jersey economic adviser has indicated that the long-term difference between general salary increases in the Jersey economy and Jersey RPI has averaged around 0.7% p.a. since 1991, with increases being higher earlier in the period and lower more recently. The States of Jersey economic adviser has also noted that public sector earnings have risen by less than private sector earnings since 1999.

In addition, we understand that general pay awards for Teachers in recent years have been substantially below Jersey RPI.

In light of this recent experience, but bearing in mind that recent experience may not be representative of the long-term future, we have assumed that general salary inflation will be equal to Jersey RPI plus 1% p.a. (i.e. a reduction of 0.5% p.a. compared with the assumption for the 2013 valuation).

Promotional salary increases

In addition to the allowance for general salary increases, an explicit age-related promotional scale was adopted at the 2013 valuation (a different scale is used for males and females).

Experience over 2014-2016 suggests that promotional increases have been higher than expected. However, we understand that recent experience may not be representative of promotional salary increases in future. We have therefore retained the scale used at the 2013 valuation.

The allowance included for promotional salary increases (in addition to general salary increases) at specimen ages is:

Expenses

Excluding investment-related expenses (which are taken into account in the net investment return assumption), we have analysed the expenses of administering the Fund during 2014- 2016 and compared this with the assumption of 1.2% of salaries adopted at the 2013 valuation. Our analysis confirmed that the 2013 valuation assumption remains appropriate.

Appendix 7: Demographic assumptions

Introduction

In this appendix, the demographic assumptions are described and we comment on how they compare with actual experience. The demographic assumptions that have been chosen are consistent with the funding target set out in the "Valuation approach" section of this report and each assumption is intended to represent a reasonable best estimate of the future.

Mortality rates before retirement

There have been only 3 deaths before retirement during 2014-2016. As there is insufficient data to carry out a credible analysis, and this assumption has little impact on the liabilities, we have retained the assumptions used for the 2013 valuation.

Specimen rates of death before retirement assumed at this valuation are set out below (per 100,000 members):

Mortality rates after retirement – current mortality rates

We have analysed the mortality experience of the Fund over the three year period from

1 January 2014 to 31 December 2016. We have used the latest set of standard tables (i.e. the SAPS S2 tables) whereas the 2013 valuation used the SAPS S1 tables.

We have set out below the ratio of actual deaths to expected deaths over the period (weighted by pension amount), with expected deaths based on 100% of the standard SAPS S2 "All lives" tables.

Mortality assumption  Males  Females

100% of SAPS S2 "All  86%  94% lives"

The experience investigation over the three years to 31 December 2016 indicates that there have been fewer deaths (by pension amount) than suggested by the standard SAPS S2 "All lives" tables. However, the small amount of experience data (66 deaths) over the 2014-2016 period means that the results of the analysis are not definitive.

We have assumed current mortality rates in line with the SAPS S2 "All lives" tables (S2PXA) with 95% scaling factor. This updates for the latest tables, but retains similar levels of current mortality as assumed in the 2013 valuation (i.e. before allowing for future improvements). Taking into account the nature of the workforce, we believe that this is a reasonable best- estimate assumption.

Mortality rates after retirement – allowance for future improvements

It is not straightforward to make an assumption about future rates of mortality improvement. In forming a best estimate assumption, we believe it is appropriate to have regard to:

• Current trends;
• Long-term trends;
• Observed generational differences, which suggest faster improvements within certain generations of pensioner (known as the cohort effect); and
• The outlook for future medical advances.

However, the allowance made must inevitably be subjective.

We have reviewed the continuing appropriateness of adopting "UK-based" improvement assumptions for a Jersey-based scheme. Our conclusion is that there is no reason not to continue using "UK-based" improvement assumptions, particularly given that Jersey population projections allow for future improvements in mortality in line with the improvements assumed for the UK population projections.

In determining an allowance for future improvements in life expectancy, it makes sense to consider the near future and longer term separately:

• Recent improvements in life expectancy are likely to be the best guide for what will happen in the near future and so improvements in the near future are best modelled by continuing recent trends.
• The forces driving longer term improvements may be very different to those behind recent improvements. This means that the assumption for long-term improvements is more subjective and should take into account analysis of historic long-term rates of improvements (and what has caused them) as well as opinions on what might happen in the future.

In November 2009, the Continuous Mortality Investigation (CMI), a group set up by the UK Actuarial Profession, published its Mortality Projections Model. The model uses complex methods for taking recent rates of mortality improvements and blending these to the long- term rate of improvements. The latest annual update of the CMI's model is known as the

'CMI_2016' projections.

CMI_2016 reflects recent data which suggests that the UK has experienced a relatively high number of deaths in the last couple of years compared to the preceding decade, the effect of which is to reduce projected life expectancies.

The responsiveness of CMI_2016 to new data can be controlled by changing the parameter S. A higher value applies more smoothing of experience data. The CMI suggest a core assumption based on the national data of S=7.5 but encourages users to consider the appropriate value for their purpose.

Our analysis suggests that the average Defined Benefit pension scheme membership does not appear to have seen such a large fall in recent longevity improvements as the national UK population. Our conclusion is that there is good reason to apply more smoothing of recent data by using S=8. This results in higher assumed life expectancies at age 65 compared with the CMI's suggested core assumption of S=7.5.

Our analysis suggests that future long-term improvements in mortality rates of between 1.0% p.a. and 2.0% p.a. for both men and women may be considered reasonable.

We have therefore assumed future improvements in mortality rates in line with the CMI_2016 Core Projections, with a smoothing parameter of S8 (which increases the assumed life expectancies) and a long-term rate of improvement of 1.5% p.a..

Retirement in normal health

We have assumed that active members will retire at the ages set out in the following table:

The analysis of the retirement experience over the period 2014-2016 for the Existing Members shows the average retirement age to be in line with the assumption adopted for the 2013 valuation.

There is only a small amount of experience data (7 retirements in normal-health) for New Members over the period 2014-2016 so we have retained the assumptions used for the 2013 valuation.

Retirement in ill-health

Over the period 2014-2016, there have been 6 ill-health retirements. As this is insufficient data for a credible analysis, we have not carried out a detailed analysis.

Experience prior to the 2013 valuation was broadly in line with the current assumptions so we have retained the assumptions used for the 2013 valuation.

Specimen rates of retirement due to ill-health assumed at this valuation are set out below (per 100,000 members):

Allowance for commutation

The 2013 valuation made allowance for New Members to commute 16.67% of their pension on retirement (there is no provision for Existing Members to commute pension for an additional lump sum).

Based on the 8 retirements from this category during 2014-2016, New Members have commuted on average about 21% of their pension on retirement. The amount of data is not sufficient for the analysis to be statistically credible but we note that the experience is in line with the experience observed for the Public Employees Contributory Retirement Scheme (PECRS) in the analysis carried out for the 2013 valuation.

We also note that the New Members Orders have been amended from 23 December 2016 to allow members in this category to exchange up to 30% of their pension for a lump sum at retirement (as opposed to 25% previously).

We have therefore assumed that New Members will commute 21% of their pension on retirement (this is equivalent to 70% of the maximum allowed). This is the assumption adopted for the 2013 valuation of PECRS, which reflected the increase in the maximum amount of pension that can be commuted. (There is much more experience of commutation rates in PECRS so this has been used as a guide in the absence of sufficient specific data for the Fund.)

Withdrawal rates

We have carried out an analysis comparing the actual number of withdrawals over 2014- 2016 with the expected number of withdrawals. Our analysis showed that the actual number of withdrawals has been lower for women but almost exactly in line with the assumption for men.

The experience analysis at the 2013 valuation suggested that experience over the period to 2013 was broadly in line with expectations for females, although withdrawals were slightly higher than expected for males.

In light of the longer term experience, we have retained the assumptions used for the 2013 valuation, except that we have simplified the assumptions such that the same withdrawal

rates are assumed for all members irrespective of service.

Specimen rates of withdrawal assumed at this valuation are as follows (per 1,000 members):

Family assumptions Family assumptions cover:

• the proportions of deaths of members and pensioners which give rise to a spouse's, civil partners' or dependant's pension;
• the age difference between the member and spouse/dependant at date of death;
• the proportions of widows and widowers who subsequently get remarried; and
• the allowance for children's pensions.

There is insufficient data to carry out a credible experience analysis for these assumptions.

On grounds of materiality, we have simplified the proportion married and age difference assumptions used for the 2013 valuation. For this valuation we have assumed that:

1. 85% of males and 70% of females are married, or have a dependant, at age 65 (or earlier death), reducing thereafter in line with spouse's mortality rates.
2. Male members are 5 years older than their dependants and female members are the same age as their dependants.

The Dedicated Pensions Unit of the States Treasury Department have advised us that, in recent times, there has only been 1 case where they have stopped a dependant's pension because of remarriage. In the light of this experience, we have removed the allowance made in the 2013 valuation for widows and widowers getting remarried.

No allowance was made for children's pensions at the 2013 valuation. For this valuation, we have therefore allowed for this by applying a loading of 10% to death before retirement

liabilities.

Re-entry to active service

The 2013 valuation allowed for deferred members to re-enter active service in future by applying a loading of 0.5% pa to the revaluation in deferment assumption. There is insufficient data to analyse the experience but this does not seem an unreasonable allowance, as our understanding is that many teachers do re-enter service and are entitled to opt for final salary linkage on their former deferred benefits (which may generate additional liabilities). We have therefore retained the assumption used for the 2013 valuation.

Appendix 8: Summary of assumptions

Financial assumptions

Demographic assumptions

Appendix 9: Discontinuance test

The discontinuance funding ratio at 31 December 2016 is 95%.

Even though the Orders governing the Fund do not envisage the Fund's discontinuance (i.e. the future accrual of benefits and payment of contributions into the Fund being discontinued), it is our practice at valuations also to review what the financial position of the Fund would have been had discontinuance occurred on the valuation date. This is done by comparing the value of the basic accrued benefits as at 31 December 2016 with the value of the Fund's existing assets at that date.

By basic accrued benefits we mean:

1. benefits in respect of current pensioners and their spouses and dependants;
2. retirement and death benefits in respect of former employees entitled to deferred pensions;
3. accrued retirement and death benefits in respect of current members based on pensionable pay at 31 December 2016, no allowance being made for pay increases after that date.

We have taken the value of the basic accrued benefits on discontinuance at the valuation date as an estimate of the terms that might be offered by insurance companies for determining the cost of immediate and deferred annuities, plus a provision to cover expenses.

In practice, if the Fund were ever to be discontinued, it is possible that the Fund would continue as a closed fund.

Derivation of assumptions

In setting the assumptions for the discontinuance test we have taken into account actual buy-out terms available in the market at the valuation date. However, we have not carried out a detailed analysis of the cost of risks that might apply specifically to the Fund and so our estimate is only a guide. Market changes to both interest rates, and demand and supply for this type of business, mean that no reliable estimate can be made, and that ultimately the actual true position can only be established by completing a buy-out.

We have set the discount rate for this estimate equal to:

• Current pensioners: Aon Hewitt Bulk Annuity Market Monitor yield curve for pensioners, which is constructed from swap and UK corporate bond market curves;
• Future pensioners: Aon Hewitt Bulk Annuity Market Monitor yield curve for non- pensioners, which is constructed from swap and UK corporate bond market curves (this applies over the period before and after retirement).

The allowance we have made for expenses is separate.

The Orders governing the Fund provide for annual increases in line with the Jersey RPI at present, although lower increases may be paid where an actuarial review has disclosed that the financial condition of the Fund is no longer satisfactory. We have assumed that in a discontinuance situation the pension increases provided would be equal to the minimum increases specified in the Orders, i.e. nil increases.

Expenses

The reserve for expenses allows for deductions to allow for the cost of forced sales of assets, an allowance for the management expenses associated with winding up and an estimate of the per member charges expected to be levied by an insurance company on buy-out.

For the purposes of disclosure in the valuation, assets are taken at their audited market value. The above allowances for expenses are therefore all presented as additions to the liabilities.

Discontinuance test results

We have considered the discontinuance position on the assumption that in the event of the Fund's discontinuance the capitalised value of the pension increase debt would be paid off in full by the States of Jersey at that point or over a period of time. This is consistent with the States of Jersey having publicly recognised a provision within its annual accounts for the debt to be settled in full, even though the mechanics of the debt repayment mechanism have not yet been formally agreed.

The results of the hypothetical discontinuance valuation are as follows:

Summary of assumptions

The table below shows the main assumptions underlying the discontinuance test, where these are different from those used for the main valuation basis.

Comparison with alternative discontinuance measure used at previous valuation

The discontinuance funding ratio calculated at previous valuations has included allowance for pension increases in line with full UK RPI increases. The discontinuance funding ratio on this alternative basis at 31 December 2016 would be 39%, which compares with a corresponding ratio of 60% at the 2013 valuation.

The value of accrued benefits on this alternative discontinuance measure has increased significantly since the last valuation. This reflects an increase in the estimated costs of buying out benefits with an insurance company due to the reduction in yields since the last valuation. The effect on this measure of the increase in the value of accrued benefits has been slightly offset by the increase in the value of assets.

Glossary

Discount rate

This is used to place a present value on a future payment. A 'risk-free' discount rate is usually derived from the investment return achievable by investing in government gilt-edged stock. A discount rate higher than the 'risk-free' rate is often used to allow for some of the extra investment return that is expected by investing in assets other than gilts.

Funding ratio

This is the ratio of the resources of the Fund (its assets, plus the value of the future pension increase debt repayments) to the resources that would be required to meet the funding target.

Funding target

This is that, based on best estimate assumptions, the assets and future contributions should be sufficient over the long term to support the benefits payable from the Fund in respect of the current members of the Fund. The resources of the Fund required to meet the funding target are determined by assessing the present value of the benefits that will be paid from the Fund in the future, based on pensionable service prior to the valuation date, plus the extent to which the present value of future service benefits for current members exceeds the present value of anticipated future service contributions for such members.

Present value

Actuarial valuations involve projections of pay, pensions and other benefits into the future. To express the value of the projected benefits in terms of a cash amount at the valuation date, the projected amounts are discounted back to the valuation date by a discount rate. This value is known as the present value. For example, if the discount rate was 6% a year and if we had to pay a lump sum of £1,060 in one year's time the present value would be £1,000.