Defined Benefit Pension Funding – Iain Clacher & Con Keating

Iain Clacher and Con Keating

Defined Benefit Pension Funding
Iain Clacher & Con Keating

The article  appears in the Securities Institute Review which is the house magazine of the Chartered Institute of Securities and Investment (CISI)
It is here 

In a range of conference presentations, articles, and commentaries, we have previously drawn attention to the fact that since 2022, there have been growing and significant differences between the estimates for the total value of assets held by DB pension schemes provided by the Pensions Regulator (TPR) and the Pension Protection Fund (PPF), which are a mixture of scheme returns and actuarial adjustments, and the Office for National Statistics (ONS) via the Financial Survey of Pension Schemes (FSPS) .

The declines in asset values reported by these bodies were extremely large and the differences in reported values were also significant. Table 1 below shows the declines in asset values between the end of 2021 and March 2023 reported by ONS, TPR and PPF, and the declines of these values, together with the proportion of these declines relative to the ONS decline.

It appears that the statistics of TPR and PPF both dramatically under stated the magnitude of the 2022 market crisis.

Table 1: Reported asset values as at December 2021 and March 2023

At March 2023, TPR reported £192 billion more in assets than ONS and PPF £181 billion.
It is worth considering the magnitude and duration of this episode. All three bodies reported minimum asset values in September 2023, Table 2 shows the minimum reported asset values and the declines over this period.

Table 2: Asset declines to lowest point reported

At this low point in asset values, the difference between TPR and PPF asset values compared to the ONS were £261 billion and £219 billion respectively. Even though asset values improved in the last quarter of 2023, the differences with the ONS figures continued to grow.
Table 3 below shows the values originally reported as of December 2023 by TPR, PPF and ONS, the last date for which we possess full data estimates

Table 3: Asset estimates December 2023

These differences are at their highest level for TPR at this time, with the gap being some £327 billion between TPR and ONS. The difference between the PPF and ONS values in March 2024 was again large, at £255 billion .

As part of the Work and Pensions Select Committee investigations into Liability Driven Investment and its consequences, the Committee recommended that these three parties work together to resolve the significant and persistent differences in the estimates of the assets held by DB schemes : “to reach an understanding of the funding position of DB schemes and publish the results”. On December 4th, 2024, a joint statement from the ONS, TPR, and PPF was published and on December 5th, 2024, the PPF published its annual Purple Book , which contains significant data revisions.

With reference to the task set by Work and Pensions, the Joint Statement states that “To help achieve this, both TPR and the PPF have amended their actuarial models to take account of benefit payments and future accrual. TPR has already made this change, which was included in its publication of analysis of the DB funding code and Fast Track. ”
It should be noted that future accrual, a liability factor, is irrelevant to the value of assets. There is also a significant problem of scale.

Based on the FSPS , total benefits paid by UK DB schemes since the beginning of 2022, when these discrepancies first became evident, until March 2024, have been £112 billion, but over this period, there have been £45.3 billion of contributions paid into schemes. In the period from the end of 2021 until March 2023, the date of PPF and TPR revisions, benefits paid were £60.4 billion and contributions amounted to £28.1 billion. The revision to TPR published asset values was from £1,415 billion to £1,284 billion, a decline of £131 billion. The revision to PPF asset values was from £1,404 billion to £1,238 billion, a decline of £166 billion.

The benefits and contribution flows are far too small to explain the revisions made. The Joint Statement must be incomplete with respect to the revisions made by TPR and PPF.

The DB Funding Code

This revision to asset values in the Funding Code Response of TPR passed without comment in the pensions trade press at the time of publication, despite these being rather large.

There is a further issue of attributing the discrepancies which emerged since 2022 to benefits payments, which is simply that they were not evident before that date. Prior to 2022 there was broad agreement across the data provided by TPR, PPF, and the ONS as to what the assets held by DB pensions were. However, if the differences that emerged after 2022 had been systematic, as omission of benefits would have been, they would have persisted consistently through time as opposed to appearing and then continuing to grow.
TPR’s revised March 2023 asset value is £1,284 billion which differs from the ONS estimate of £1,223 billion, a difference of £61 billion, which is 5%, and £61 billion is a material amount.

The Purple Book is based upon the revised figures of TPR but uses their own roll-forward algorithms. However, while there is much closer agreement among the asset values reported by ONS, TPR and PPF, as shown in table 4 below, there is still cause for concern. Figure 1 shows the linear regression of PPF asset values on the ONS figures.

The explanatory power of the model,(R2 ) is low at 78% and the coefficient of 1.049, implies that the PPF figures are high biased and will, other than for trivially small values of assets, produce high estimates relative to the ONS. However, this overestimate will be small relative to those seen recently – at the highpoint of values the end of 2021 (£1,821 billion) the expected overestimate produced by PPF would be £49.7 billion. We will be updating this model and publishing the results as more data become available.

Table 4: Revised asset values

Figure 1: Regression of revised PPF values on ONS data

DB Liability Estimates

In looking at estimates of the liabilities of DB pension funds, TPR’s revision of liabilities was markedly different from that of the PPF, as Table 5 below shows. TPR report values for Technical Provisions (TPs) liabilities, while the PPF reports the Section 179 valuation of the liabilities, which arise from the reduced benefits that PPF pays to members of schemes that end up in the PPF.

First, the revisions to liabilities are wildly different. TPR’s TP liabilities decline by £96 billion (8%) but PPF’s S179 liabilities decline by only £14 billion (1.3%). Second, and more importantly, there is an inconsistency between these two amounts. PPF benefits were 13.1% lower than TPs in the original estimates prior to the revisions but are only 6.8% lower after the revisions.

Given the PPF’s rules and the demographics of schemes in the PPF universe, this latter cost to potential PPF members seems highly improbable.

Table 5: TPR and PPF liability estimates

Scheme Funding

On scheme funding, the Joint Statement asserts that:

“Because of the above factors, it is currently not appropriate to compare one set of statistics from one organisation with another’s set of statistics in isolation because these are not like for like. Instead, when considering funding levels, or surpluses and deficits, only TPR and the PPF statistics should be used as these are calculated consistently within each organisation. For example, it is not appropriate to compare liabilities estimated by TPR or the PPF with ONS FSPS assets.”

This seems like a rather odd statement. The task set by Parliament was to work out why the differences were so large between FSPS asset values, and the statistics produced by TPR and PPF. Given that these differences have now been largely resolved, and it seems that on balance the statistics of the ONS have proven to better reflect the turmoil and impact of the LDI crisis, then surely a combination of ONS assets and TPR liabilities is perfectly reasonable.

On a more positive note, PPF’s revision to full buy-out values corrects the highly questionable value (highlighted in purple) reported in the 2023 Purple Book – see Table 4.

Table 4: Full Buy-Out values and funding ratios 2022 – 2024

Final thoughts

Despite all of this, we still do not know the true cost of LDI or what assets were sold to meet the significant collateral calls faced by schemes. However, it is worth noting that the ONS, via the FSPS, collects but does not publish data on the purchases and sales of investments. This data could be used to resolve unequivocally the matter of the losses incurred by schemes due to LDI, and not only how much was lost, but what was sold to meet margin calls.

We can offer some insight using the FSPS data on holdings of pooled LDI funds, gilts and corporate bonds. Figure 2 shows the proportional changes in these asset holdings by quarter.

Figure 2: Proportional Losses and Gains 2022-2024

As expected, pooled LDI funds experienced larger losses (as leveraged instruments) than gilts or corporate bonds in Q1 and Q2 2022. However, in Q3 pooled LDI funds declined less than either gilts or corporates, which may be interpreted as being due to payments (collateral calls) being made into these funds. These payments into pooled LDI funds would have been of the order of £15-£20 billion pounds. The fourth quarter shows unequivocally flows into pooled funds as the equity in these rises by £28 billion.

If this payment was met by sales of gilts and corporates, it would account for some 60% of the declines evident, and total transfers of £32 – £37 billion to pooled LDI funds would restore the expected performance relationship. One further observation is that, in the post crisis period, pooled LDI funds did not in general deliver the upside performance expected of such levered funds.

 

About henry tapper

Founder of the Pension PlayPen,, partner of Stella, father of Olly . I am the Pension Plowman
This entry was posted in pensions and tagged , , , , , , . Bookmark the permalink.

Leave a Reply