Valuations and recovery plans of UK defined benefit (DB) and hybrid pension schemes.

Published: 11 August 2020

## Overview

This is the 2020 update to The Pensions Regulator’s (TPR) annual funding statistics for UK defined benefit (DB) and hybrid schemes. The tables in the Annexure are prepared in accordance with the UK code of practice for official statistics. The underlying data are sourced from valuations and recovery plans submitted to TPR by schemes with deficit positions, and from annual scheme returns for schemes with surplus positions.

The update is based on Tranche 13 schemes (with effective valuation dates falling from 22 September 2017 to 21 September 2018 inclusive). These valuations (with due dates for receipt falling within the period December 2018 to December 2019) fall within the fifth triennial cycle of scheme funding.

The report shows funding trends in the context of market conditions, assumptions and scheme characteristics that impact on valuations. It also describes reported arrangements for recovery plans, employer contributions and contingent security.

The bulletin is comprised of several sections. Specifically, Section 2 summarises the key figures in the current tranche. Section 3 discusses market conditions. Section 4 summarises funding levels and Section 5 covers various aspects of funding strategies. Section 6 reports valuation assumptions. Data summaries (tables) contained in the Annexure provide more detail on the high-level trends presented in this document. Note that Tranche 13 belongs in the cohort with Tranches 1/4/7/10, meaning comparisons with other years will generally be limited to these tranches.

## Tranche 13 summary

By the end of January 2020, TPR had received over 1,730 valuations (Annexure Table 1.1) with an effective valuation date for Tranche 13 - over one third (37%; Annexure Table 1.2a) of which reported a surplus on the technical provisions (TPs) funding basis. Of schemes submitting these valuations, 74.8% had previously submitted valuations in Tranches 10, 7, 4, and 1.

Assets and deficits in this bulletin are defined on a Section 75 (s75) buyout basis. See the Glossary section in the Annexure for further information. The growth in assets exceeded the growth in liabilities between Tranche 10 and 13 valuation dates for a majority of schemes, resulting in a relatively increased average^{[1]} funding ratio on a TPs basis (Annexure Table 2.1a). Compared to Tranche 10, average annual deficit reduction contributions (DRCs) as a proportion of TPs for Tranche 13 are relatively unchanged, at the median in nominal terms (Annexure Table 3.4). The relative increase in average annual DRCs between Tranche 10 and 13 is 3% at the median, while the median extension to the recovery plan end date is two years.

Figure 1, below, illustrates the distribution of changes in deficits measured on a TPs basis for Tranche 13 schemes that previously submitted Tranche 10 valuations. At the median, scheme assets grew by 17%, whereas TPs grew by 10%. The median relative change in deficits for all schemes is -33%, with around 70% of all schemes experiencing a reduction in deficit over the inter-valuation period.

### Figure 1: Distribution of the change in deficits from Tranche 10 to Tranche 13 (all schemes in both tranches)

### Key figures at a glance

- 37% of schemes in surplus (Annexure Table 1.2a).
- 93.4% (95.1%): average (median) ratio of assets to TPs for schemes in deficit and surplus (Annexure Table 2.1a).
- 109.7% (106.6%): average (median) ratio of assets to TPs for schemes in surplus (Annexure Table 2.3a).
- 84.1% (87.2%): average (median) ratio of assets to TPs for schemes in deficit (Annexure Table 2.3a).
- 6.1 (5.2) years: average (median) RP length for schemes in deficit (Annexure Table 3.1).
- 2.62% (2.61%): average (median) nominal single effective discount rate (SEDR) for Tranche 13 schemes in deficit and surplus; with 50% of assumptions falling on or between 2.2% and 3% (Annexure Table 4.3). SEDR is defined in the Methodology section of the Annexure.
- -0.77% (-0.77%): average (median) real SEDR for schemes in deficit and surplus; with 50% of assumptions falling on or between -1.16% and -0.39% (Annexure Table 4.2).
- 89 (91) years: median life expectancy of a future male (female) pensioner currently aged 45 for schemes in deficit and surplus (Annexure Tables 6.1a & 6.1c).

## Market conditions

Equity markets performed positively during the inter-valuation period (Tranche 10 to Tranche 13 valuations). Relative to 31 March 2018 (the effective date of many Tranche 13 valuations), the FTSE All-Share Index realised a three-year total return of 18.6%^{[2]}.

Between Tranche 10 and Tranche 13 valuations, yields on corporate and government bonds continued a declining trend, falling to a trough in Q2 of financial year 2016 following the Brexit referendum, before rising gradually.

The real 20-year spot rate of interest, already in the region of negative yields during the previous valuation period, remained negative (-1.6%) towards the end of the Tranche 13 valuation period.

Figure 2, below, shows market expectations for real interest rates as estimated by the Bank of England in the inter-valuation period. Forward interest rates at the end of March 2018 show an expectation for lower yields over all maturities relative to the position a year earlier and in March 2015.

### Figure 2: UK instantaneous real forward curves

## Funding levels

Figure 3, below, reports the distribution of the ratio of assets to TPs liabilities for Tranche 13 schemes, by scheme maturity group (four categories describing a scheme’s ratio of pensioner TPs to total TPs).

The median level of TP funding generally increases across each of the maturity groups, from 93% among schemes with maturity levels less than 25%, to 102% for schemes with maturity levels of 75% or greater.

### Figure 3: Assets to Part 3 (TP) Liabilities (per cent) by maturity (Tranche 13, all schemes)

Tranche 12 schemes saw improved asset positions. This trend has continued for Tranche 13 schemes, which may have been due to the combined impact of:

- sponsor contributions (regular and remedial), and
- positive (overall) gains on investments in the three years to valuation

Liabilities have also grown due, in part, to lower investment return assumptions relative to those assumed in the previous valuation - which in turn may have been driven by revised (lower) expectations for long-dated bond yields relative to Tranche 10 valuations. In general, schemes with hedged positions in respect of interest rate risk will have fared better overall.

- 93.4% (95.1%): average (median) ratio of assets to TPs for schemes in deficit and surplus (Annexure Table 2.1a).
- 109.7% (106.6%): average (median) ratio of assets to TPs for schemes in surplus (Annexure Table 2.3a).
- 84.1% (87.2%): average (median) ratio of assets to TPs for schemes in deficit (Annexure Table 2.3a).

Figure 4, below, illustrates the distribution of the ratio of assets to TPs for Tranches 1 - 13. Compared to all other tranches, funding levels in Tranche 13 are higher across all percentiles of the distribution. At the median, the TP funding ratio for Tranche 13 is 95%, compared with 90% among the Tranche 10 cohort.

### Figure 4: Ratio of assets to Part 3 (TP) Liabilities (per cent) (all schemes, all tranches)

For schemes submitting valuations in respect of both Tranche 10 and Tranche 13, the median increase in assets between valuations is 17%, and the median increase is 10% for TPs. The average ratio of assets to TPs for Tranche 13 is 93.4% (96.4% weighted^{[3]}), a 4.8 percentage point increase on average over Tranche 10. The average ratio is generally higher for schemes:

- that report liabilities in respect of active memberships
- with stronger covenants (deficit schemes only)
- without a contingent asset
- with shorter recovery plans

See Tables 2.1 and 2.3 in the Annexure for further detail on funding ratios

For schemes in surplus only, the average ratio of assets to TPs is 109.7% (106.7% weighted).

Relatively higher funding among schemes reporting TPs in respect of active members could reflect that less well-funded schemes are more likely to close to future accruals.

## Funding strategies

### Recovery plans

Approximately one-fifth (20%) of schemes with valuations in both Tranche 10 and Tranche 13, reported a surplus of assets over liabilities in both tranches. Around one in six schemes (17%) moved into a position of surplus, with less than one in thirty (2.7%) moving into deficit, from a surplus position in Tranche 10.

Therefore there are around three-fifths (60%), of schemes with valuations in both Tranche 10 and Tranche 13 which reported a shortfall of assets against liabilities in both tranches. For around three-quarters (76%) of these schemes, a reduction in the size of the deficit between tranches (amongst other factors), has led to a shortened recovery plan period in Tranche 13, relative to that agreed under the Tranche 10 valuation.

Given the (broadly) three year inter-valuation period itself however, a reduction in recovery plan length of less than three years between tranches still represents an extension to the date at which the scheme is anticipated to be fully funded. Figure 5, below, shows the distribution of changes to recovery plan end dates for Tranche 13 schemes that had previously submitted Tranche 10 valuations. Just over two thirds of schemes (68%) - including those which remain in surplus - have brought forward their recovery plan end dates or have left them broadly unchanged. Around one in six (17%) schemes have extended their recovery plan end-date by up to three years (representing a recovery plan of a similar length in Tranche 13 relative to that agreed in Tranche 10), with a similar proportion (15%) extending their recovery plan end-date by more than three years.

### Figure 5: Distribution of changes to recovery plan end dates, by Tranche 10 and Tranche 13 recovery plan length groups (schemes in both Tranches 10 and 13)

Figure 6, below, highlights the distribution of recovery plan end dates by Tranche. Although the recovery plan end-dates for recovery plans has increased in Tranche 13 across the distribution, there has been a narrowing at the upper tail of the distribution, ie 75th percentile and above.

Compared with Tranche 10, Tranche 13 end dates have increased by two years at the median, 1.1 years at the 75th percentile and 0.4 years at the 95th percentile.

### Figure 6: Distribution of recovery plan end-dates (schemes in deficit, all tranches)

In respect of their first valuations under scheme specific funding, schemes in Tranche 1 had a median recovery plan end date falling in 2016. Under the fifth cycle of funding, broadly the same set of schemes have a median recovery plan end date falling in 2024.

The mean and median recovery plan lengths for Tranche 13 schemes in deficit are 6.1 and 5.2 years respectively. For Tranche 10 these figures were 7.3 and 6.4 years respectively (Annexure Tables 3.1 and 3.3).

Longer recovery plans tend to be associated with schemes:

- with weaker covenant support
^{[4]} - with lower levels of funding on a TP basis
- with at least one contingent asset (Annexure Table 3.3)

See Tables 3.1, 3.2 and 3.3 in the Annexure for more detail on recovery plan length

Figure 7, below, shows the distribution of recovery plan lengths for Tranche 13 schemes by covenant group ^{[5]}. Three-quarters of schemes in covenant group 1 (strong) have recovery plans of up to around 6.7 years, while the same proportion of schemes in covenant group 4 (weak) have recovery plans of up to around 12.2 years.

### Figure 7: Distribution of recovery plan length by covenant group (Tranche 13 schemes in deficit)

The table, below, shows the median extension to recovery plan end date extension according to covenant group migration. The largest increase is 1.4 years, for schemes migrating to covenant group 4 from covenant group 2 over the inter-valuation period. Generally, a strengthening in covenant for initally weaker schemes is associated with a smaller extension to recovery plans, compared with that under a static transition or a weakening of covenant.

### Table: Median recovery plan end date extension by covenant group migration (Tranche 13 schemes in deficit only)

Covenant Group (Tranche 13) | ||||
---|---|---|---|---|

Covenant Group (Tranche 10) | 1 | 2 | 3 | 4 |

1 | 0.7 | 0.3 | ||

2 | 0.1 | 0.3 | 0.7 | 1.4 |

3 | 0.0 | 0.0 | 0.0 | 1.0 |

4 | -0.3 | 1.0 |

Table: Median recovery plan end date extension by covenant group migration

Source(s): TPR

### Contributions

As a proportion of liabilities calculated on a TPs basis, average annual DRCs for Tranche 13 schemes are unchanged at 2.2%, relative to Tranche 10.

This is influenced by increases in liabilities calculated on a TPs basis, as well as changes in nominal DRCs. The median increase in TPs is 10% and the corresponding relative increase in average annual DRCs is 3%. A higher level of DRCs as a percentage of liabilities on a TPs basis is associated with:

- shorter recovery plans
- smaller schemes (by both TPs and members)
- lower maturity (schemes with pensioner liabilities that are less than 25% of total liabilities)
- schemes with a lower ratio of assets to TPs
- schemes with a less than 20% allocation to return seeking assets

Figure 8, below, shows the extent to which the percentage point increase in average annual DRCs as a proportion of TPs, varies by the percentage point increase in TPs funding for Tranche 13 schemes in deficit. It can be seen that the majority of schemes with positive changes in funding ratios from Tranche 10 to 13 have changes to average annual DRCs as a proportion of liabilities, that are concentrated in the range: -1% to 0.5%.

Schemes with negative changes in the funding ratio have changes to the average annual DRCs as a proportion of liabilities, that are concentrated in the range: -0.5 to 1%. The level of DRCs will be influenced by extensions made to recovery plan end dates (two years at the median) as well as changes to the size of deficits, amongst other factors.

### Figure 8: Absolute percentage increase in DRCs as a proportion of TP liabilities vs. Absolute percentage increase in the ratio of assets to TPs (schemes in deficit only)

### Asset Allocation

The majority of Tranche 13 schemes have less than 60% of assets invested in return-seeking^{[6]} asset classes, while less than a fifth of schemes have a higher allocation (>60%) to return-seeking asset classes (Annexure Table 1.2a).

Among Tranche 13 schemes there is no evidence of a lower prevalence of riskier investments among schemes with a stronger covenant. This could be because schemes with stronger covenants tend to be better funded, and better funded schemes may be more likely to afford (and hence adopt) de-risking strategies.

Figure 9a, below, shows the distribution of return-seeking assets by covenant group. The median allocation to return-seeking assets is 46% for covenant group 2 schemes, compared to 48% for covenant group 4 schemes. Generally we observe an increased allocation to return-seeking assets with declining covenant strength.

### Figure 9a: Distribution of return-seeking assets by covenant group (Tranche 13 schemes in deficit)

Figure 9b, below, shows the distribution of return-seeking assets by maturity group (as measured by the ratio of pensioner liabilities to total liabilities). The median allocation to return-seeking assets is 48% for schemes with 25% to less than 50% maturity, compared to 36% for schemes with greater than 75% maturity. Generally we observe a decreased allocation to return-seeking assets with increasing maturity, though this relationship does not hold for the most immature schemes.

### Figure 9b: Distribution of return-seeking assets by scheme maturity group (Tranche 13 schemes in deficit)

### Contingent security

Less than a fifth (18%) of Tranche 13 schemes have additional security in the form of one or more contingent assets, which typically, but not always, take the form of guarantees from a sponsor’s parent or associated group entity.

Around a tenth (8.8%) of schemes have contingent assets that are formally recognised by the PPF in the calculation of the PPF risk-based levy. A similar proportion (9.4%) have contingent assets that are not recognised by the PPF but are reported as additional security in support of funding.

The presence of contingent assets is associated with:

- larger schemes (by both members and liabilities)
- weaker covenant support
- longer recovery plans

See Tables 2.6 and 2.7 in the Annexure for more information on contingent security

## Valuation assumptions

### Discount rates

Market experience and changing expectations over the inter-valuation period may be reflected in discount rates to varying degrees - depending on investment strategy and/or valuation approach. The statistics show an association between the TP’s discount rate and a scheme’s investment strategy. (The investment strategy of a scheme is approximated here by the scheme’s allocation to return-seeking assets and may itself take some account of the ability of the employer to underwrite downside risk).

Between Tranche 10 and Tranche 13 valuations, the 20-year nominal spot rate on gilts declined overall having reached an historic low of 1.34% in August 2016 before rising to 1.92% at the end of the Tranche 13 valuation period. The gap between implied inflation and the nominal gilt yield, which had widened during the Tranche 10 to Tranche 13 inter-valuation period, kept real yields negative over the three-year period between most valuations.

For 31 December 2017 and 31 March 2018 (the two most common Tranche 13 valuation dates), the 20-year spot rates on gilts are 1.91% and 1.76% respectively, while the corresponding median single effective discount rates (SEDR) for valuations are 2.51% and 2.6% respectively. The median SEDR fell below the rate of 20-year market implied inflation, relative to the position a year earlier.

Figure 10, below, shows the median SEDR relative to 20-year UK gilts, 20-year spot inflation, and greater than 15-year AA-rated corporate bonds (for schemes in deficit in Tranches 1 to 7 and for all schemes in Tranches 8 to 13).

### Figure 10: Median (nominal) SEDR, Bank of England nominal gilts 20-year spot rate, greater than 15-year AA corporate bonds, and BOE 20-year inflation

The average real SEDR is -0.77% for Tranche 13, compared to 0.15% for Tranche 10.

Schemes with a higher proportion of return-seeking assets tend, on average, to assume a higher discount rate. The implied level of outperformance over gilts in the discount rate appears to have an even stronger relationship to the proportion of return-seeking assets held by schemes - although there is a wide range of assumptions in discount rate outperformance among schemes with a similar allocation to return-seeking assets.

The average assumed return over nominal gilts is lower for Tranche 13, compared to Tranche 10. However, the average outperformance of the real SEDR over the 20-year real government spot rate is 0.88% for Tranche 13, compared to 0.98% for Tranche 10.

See Table 4.2 in the Annexure for more information on discount rates

### Life expectancies

While there were notable increases in average assumed life expectancies over the first three funding cycles, reflecting stronger mortality assumptions over that period, assumptions regarding average life expectancy are generally lower for Tranche 13 schemes relative to Tranches 10 and 7, for both future and current pensioners.

See Tables 6.1 - 6.2 in the Annexure for more information on life expectancies

This reflects the general trend observed in the wider population over the last few years and that almost all schemes use the CMI mortality projection model when allowing for future improvements (see Tables 5.1 - 5.4 in the Annexure).

In respect of the underlying mortality assumptions:

- 93% of Tranche 13 schemes use the Self-Administered Pension Scheme (SAPS) tables
- 55% apply a scaling factor or rating to base tables to adjust for scheme experience
- 96% use the continuous mortality investigation (CMI) projection model (first published in 2009) to allow for future improvements
- while 82% of schemes assume a long-term rate of improvement/underpin of 1.5% or higher with 5% assuming a rate of 2% or higher

Figures 11a and 11b, below, show the distribution of assumed life expectancies for future male and female pensioners currently aged 45 respectively. At the median, the assumed life expectancy of a future male pensioner currently aged 45 for Tranches 10 and 13 is 89.8 and 89 years, respectively. Similarly, the median life expectancy of a future female pensioner currently aged 45 for Tranches 10 and 13 is 92 and 91 years, respectively. The median life expectancy assumption for future pensioner males (females) in Tranche 13 is higher than the corresponding assumption for Tranche 1 by around 2 years (1.2 years).

### Figure 11a: Distribution of life expectancy assumptions for future male pensioners currently aged 45 (all tranches)

### Figure 11b: Distribution of life expectancy assumptions for future female pensioners currently aged 45 (all tranches)

## Footnotes

- [1] Averages are unweighted unless stated otherwise.
- [2] FTSE Russell Factsheet March 2018.
- [3] Weighted by TPs.
- [4] Covenant Groups (1-4) are assigned at the point of initial RP reviews to facilitate prioritisation. These grades may vary to the view taken during case-level intervention, where a wider range of information is taken into account. They are defined as: Covenant Group 1 - strong; Covenant Group 2 - tending to strong; Covenant Group 3 - tending to weak; and, Covenant Group 4 - weak. Covenant assessments are not usually undertaken for schemes in surplus.
- [5] See previous footnote.
- [6] ‘Return-seeking assets’ in this report include equities, commodities, 60% of insurance policies, 75% of property, 80% of hedge funds, 25% of corporate bonds and assets held in the ‘other’ category.