Triennial valuation 2016 in the LGPS (England & Wales) - Detail Report

Executive summary

On 31 March 2017, 91 Funds from England & Wales participating in the Local Government Pension Scheme (LGPS) submitted the results of the 2016 actuarial valuation with the purpose of setting employer contribution rates in each Fund for the period from 1 April 2017 to 31 March 2020.

This report outlines some of the headline results as well as some of the issues surrounding the 2016 valuation understood to have been faced by all the actuarial firms.

Headline results

The main reasons for any improvements tended to be good investment performance as well as the level of deficit contributions paid over the period. The intervaluation period had also seen a period of low inflation and lower future inflation expectations which resulted in a reduction in liability on the cashflow projections.

Headline issues

The LGPS Regulations for actuarial valuations have changed since the 2013 valuation and so has the context surrounding the triennial valuation. There is greater scrutiny of Funds’ strategies and approaches for example:

In addition, the 2016 valuation was the first full valuation following the change of the LGPS to a career average scheme. Other challenges included increasing numbers of employers and the implementation of the new universal data capture (UDC) specification which forms the basis of the membership data that is provided to actuaries to complete the valuation.

Introduction

This report summarises the data collected from 89 of the 91 Funds in England & Wales that participate in the LGPS (the two Environment Agency Funds are excluded for the purposes of this analysis1). It compares the 2016 results to the 2013 valuation as well as looking at the assumptions used, the investment strategies and the ranges of recovery periods and recovery plans.

In carrying out this research, the Scheme Advisory Board secretariat was supported by Barnett Waddingham in reviewing and analysing the valuation reports for all LGPS Funds E&W.

Participating employers are being under increasing cost pressures and on the whole, it has been reported that they are becoming more engaged with the valuation process. This may be due to increased efforts on the part of administering authorities. An increasing number of Funds undertook employer covenant reviews as part of the valuation process. This helps Funds to look at the appropriateness of the contributions being set and the integration of employer covenant risk, investment strategy risk and funding risk in a similar way to what is seen in the private sector.

Background

In 2016, the aggregate LGPS deficit for the 91 Funds analysed decreased from £47.5bn to £37.7bn with 79 of the Funds seeing an improvement in their funding position. This was primarily due to good investment returns as well as the level of deficit contributions paid over the period. The intervaluation period also saw a period of low inflation and lower future inflation expectations which placed a lower liability on the cashflow projections.

Overall this is a positive outcome for the LGPS although the total published deficit is still almost £38bn. Employers have committed to pay at least £2.2bn, £1.9bn and £2.0bn for the 2017/18, 2018/19 and 2019/20 Scheme years respectively. The higher amount in year one is due to some employers having a preference to front end load their contributions.

Since the valuation date, the 2016/17 year has seen some very high investment returns although this will have been offset by the worsening in market conditions (and therefore the funding assumptions) as a result of that. Future inflation projections have reversed the 2013/2016 downward trend and increased back to 2013 levels – most likely a function of the Brexit referendum outcome.

This report explores some of the key findings from looking at all the valuation reports and is split into the following sections:

Please contact the Secretariat for further information on the results of this research.

Results of analysis

Data

Membership

The LGPS continues to increase in size with total membership increasing from 4.9m members in 2013 to 5.6m in 2016. During the intervaluation period, auto-enrolment probably explains some if not all of this increase. It is also known that individuals may have multiple membership records due to different contracts etc. To confirm we have analysed the number of records and not the number of individuals.

membership 2013membership 2016

In the table below we set out how the membership has changed over the intervaluation period.

Membership Summary Number   Total Pensionable pay/pension amounts   Average Age  
  2016 2013 2016 2013 2016 2013
  000s 000s £ms £ms    
Active 1,875 1,706 31,807 26,494 48.7 48.7
Deferred 2,149 1,748 2,911 2,095 48.6 48.4
Pensioners 1,546 1,398 7,493 5,855 69.5 68.6
Total 5,570 4,852        

 

Actuarial adviser

In terms of market share, the graph below shows the proportion of Funds by number of Funds advised by each actuarial firm.

actuarial firms 2016 

Financial assumptions

This section outlines findings split between the different key assumptions which are:

Discount rate

The discount rate assumption is the key assumption driving the value of each Fund’s liabilities  To determine the value of accrued liabilities and future contribution requirements at any given point in time it is necessary to discount future payments to and from the Fund.  There are a number of different approaches which can be adopted in deriving the discount rate to be used and the approach that is most appropriate will depend on the purpose of the valuation and the overall funding objectives and risk appetite of the Fund.  Therefore each actuary and Fund will derive their assumptions in an appropriate way.

The discount rate varies between employers and employer groups in each fund, reflecting covenant.  It is important to note that analysis here is of the average discount rate for each fund.

The average disclosed discount rate used in the valuation fell from 5.1% p.a. to 4.4% p.a. and spanned a range of 3.8% to 5.7% p.a. compared with 4.5% to 6.6% p.a. in 2013.  All else being equal, a fall in the discount rate increases the value of the Scheme’s liabilities.

The chart below shows the spread of discount rate used as at 31 March 2016.

 

discount rate 2016

Inflation

In the LGPS, the Consumer Prices Index (CPI) is the current inflation measure used to index pensions in payment and deferment, and to revalue members’ CARE benefits for service after 31 March 2014. The increase is set out each year in the public service pensions increase order issued by HM Treasury.

The average disclosed CPI inflation assumption used fell from 2.6% p.a. in 2013 to 2.2% p.a. in 2016. All else being equal, this would have led to a decrease in the value of the Scheme’s liabilities.

As there is no measure for expectations in the assumption for CPI, a deduction is generally made to the Retail Price Index (RPI) inflation assumption. All actuarial firms applied a consistent difference in the RPI and CPI assumption to almost all of their Funds (some exceptions were made) but this difference did vary by actuary. Barnett Waddingham used a deduction of 0.9% p.a. whereas Mercer and Hymans Robertson used a deduction of 1.0% p.a., and Aon Hewitt, 1.1% p.a.

Inflation assumption 31 March 2016   31 March 2013  
  CPI Inflation Deduction CPI Inflation Deduction
  p.a. p.a. p.a. p.a.
Mercer 2.2% 1.0% 2.6% 1.0%
Barnett Waddingham 2.4% 0.9% 2.7% 0.8%
Hymans Robertson 2.1% 1.0% 2.5% 0.8%
Aon Hewitt 2.0% 1.1% 2.4% 0.9%

Real discount rate

The relationship between the adopted inflation assumption and discount rate assumption, the real discount rate, is arguably the most important financial assumption. For example a 0.1% increase in the inflation assumption combined with a 0.1% increase in the discount rate assumption results in very little change to the estimated value of the liabilities but a 0.1% increase in the inflation assumption combined with a 0.1% decrease in the discount rate assumption, i.e. a fall in the real discount rate of 0.2%, could lead to an increase in liabilities of around 4%.

The graph below shows the change in real discount rate assumptions, but instead focusing on the change in the real discount rate assumption. There were 11 Funds where there was no change to the real discount used and a further 26 Funds where the change was less than 0.2%. Although the picture is similar, it has a different meaning when looking at the real discount rate., all things being equal you would expect a fall in the real discount rate to result in a deterioration in the funding position , which does not correlate with the results in the graph below which shows a number of falling yields along with improved funding levels.

Therefore this means that the gains that Funds have experienced have come from other factors which could include investment performance, contributions paid, different membership experience and/or other assumption changes.

change in real discount rate

Salary increases

As the LGPS is now a CARE scheme the benefits earned after 31 March 2014 are revalued with inflation rather than salary increases, the overall effect on valuation results of the salary increase assumption is less than it has been previously and will not affect the contribution calculations. However, it still affects all pre-2014 active liabilities.

Actual salary increases in the public sector have generally been low over the past few years and at both the 2013 and 2016 valuations, there has been more of a range of salary increase assumptions used to reflect the circumstances of the individual Fund.

In 2013, all the firms included a short term salary allowance which varied by Fund in terms of both what the increase was and for how long it should apply.

In 2016, Barnett Waddingham and Mercer have continued to adopt the short-term salary overlay and in addition Barnett Waddingham lowered their expectation of long-term salary increases from 1.8% p.a. above CPI inflation to 1.5% p.a. above CPI inflation which was the same assumption used by Mercer and most of Aon Hewitt. Hymans Robertson took a different approach and instead of adopting a short-term overlay, their general expectations for future salary increases came down for all Funds.

short term salary

In addition, a promotional salary increase was included in the 2016 assumptions which is in addition to the short-term and long-term salary increases. This was used by all the actuarial firms except for Mercer.

The chart below shows the range of long-term salary increase assumptions used by Funds in the 2016 valuation.

salary increase

Demographic assumptions

Mortality

The key demographic assumption required for determining the pension liabilities is the mortality assumption. The mortality assumption will generally vary from Fund to Fund as the mortality assumption will take into account the specific profile of each Fund and its members.

There are two aspects to consider in determining appropriate post-retirement mortality assumptions:

  1. 1. choosing an appropriate mortality base table assumption applicable today taking into account characteristics of the Fund members
  2. 2. making an appropriate allowance for mortality to improve in future.

The base table will generally vary between Funds but the allowance for mortality to improve in the future is a more subjective assumption which will tend to be consistent between Funds, although the different actuarial firms have taken different views on what this should be and which version of the CMI Model to use.

The assumptions used have resulted in a range in life expectancy assumptions as set out in the table below.

Life expectancy Minimum Maximum Average
Male age 65 20.6 27.7 22.7
Female age 65 22.7 28.3 25.0

The range is significant reflecting to a large extent the quite different rates of mortality across England and Wales.

Funding

The results of the 2016 valuation reported assets of £216,600m and liabilities on local funding assumptions of £254,300m revealed a deficit of £37,700m and a funding level of 85%. This is an overall improvement compared to the position in 2013 which showed a funding level of 79% and a total deficit of £47,500m.showed an average funding level of 79%. Published funding levels ranged from 55% to 103% with an average of 85%. The Fund with the highest reported funding position was Kensington and Chelsea at 103% funded and the lowest reported funding level was LB Brent at 55% funded . The graph below shows the range of funding levels and the number of Funds that fall into each bracket.

funding level

The change in funding positions has ranged from a fall of 6% (LB of Barnet) to an improvement of 17% (North Yorkshire). 12 Funds have seen a worsening or no improvement in their funding positions with the rest seeing an overall improvement.

The graph below shows the relationship between the funding level and the real discount rate. As you can see there is a vaguely upward trend to suggest that as the real discount rate increases, the funding level improves but it is not as clear a trend as we would have expected to see.

funding level discount rate

It should be noted that these funding positions are calculated based on local funding assumptions rather than a like-for-like basis. Funds will use different assumptions to reflect their individual circumstances and attitudes to risk. However, as part of the 2016 valuation Funds were asked to submit results on a set of agreed assumptions to the Scheme Advisory Board (SAB) standard basis. This revealed an average funding level of 96% on the SAB basis. Further analysis is expected in the Section 13 report to be published by GAD.

Contributions

The purpose of the 2016 actuarial valuation was to set appropriate contribution rates for each employer in the Scheme for the period from 1 April 2017 to 31 March 2020. This is required under Regulation 62 of the LGPS Regulations. The Regulations for actuarial valuations have changed since the 2013 valuation and so has the context surrounding the valuation. Regulation 62 specifies four requirements that the actuary “must have regard to” and are detailed below:

The primary rate is the employer’s share of the cost of benefits accruing in each of the three years beginning 1 April 2017. In addition each employer pays a secondary contribution as required under Regulation 62(7) that when combined with the primary rate results in the minimum total contributions certified for each employer. This secondary rate is based on their particular circumstances and so individual adjustments are made for each employer.

Analysis has been completed on the expected level of contributions to be paid into the Fund in the 2017/18, 2018/19 and 2019/20 Scheme years based on the information included in the valuation reports.

The table below shows the range of the primary rate at whole Fund level:

primary rate

This is the weighted average across all employers within each Fund – individual employer primary rates will exhibit greater variability.

In terms of secondary contributions, the total expected value of contributions to be paid into each Fund in the three years for the period from 1 April 2017 to 31 March 2020 has been reviewed.

There is not an even spread of contributions during the three years but it is assumed that this is because some employers choose to front-end load their contributions and pay them in the first intervaluation year.

secondary contributions

In terms of recording these contributions, they were set out in the Rates and Adjustment Certificate for each Fund and the layout of the certificate was agreed between the four actuarial firms for consistency following comments from GAD on the section 13 dry run report.

Recovery plan

Where a shortfall is revealed, each employer must agree a “recovery plan” with the Administering Authority to address meeting any deficit over a “recovery period”. The actuary must certify in the Rates and Adjustments Certificate that the contributions set are sufficient to ensure that a funding level of 100% of liabilities is met by the end of the recovery period.

There is limited information in the valuation reports about the length of recovery periods used by Funds and participating employers. Some Funds have reported an average recovery period and some have reported the longest agreed recovery period for their employers (and sometimes both). However, it is not necessarily appropriate to consider a recovery period at Fund level for a number of reasons including:

In addition, recovery periods are used by some Funds as a tool to reflect employer covenant and therefore different recovery periods are offered to different employers but this will be set out in the Funding Strategy Statement of the relevant Fund.

The information disclosed in the valuation reports do not provide enough information to easily identify how recovery periods have changed since 2013 as some disclosures have changed from average period to maximum period and not all actuaries included any detail on recovery periods in their 2016 valuation reports. However using the information disclosed, it suggests that recovery periods have reduced by approximately two years on average.

The graph below shows the deficit recovery periods for 51 Funds based on that reported as at 31 March 2016:

reported recovery

Investment strategy

The proportion of assets held in each the main asset classes is shown in the chart below using the information provided in the valuation reports. For some asset classes an allocation has been made to a class which is believed to be appropriate. This is set out in the chart below:

 

investment strategy

This is roughly an 80/20 split between growth and protection assets classing equities, pooled investment vehicles, property and infrastructure as growth assets.

Please contact the Scheme Advisory Board secretariat for further information on the results of this research.

Appendix 1

List of LGPS Fund reports analysed