Office for National Statistics (ONS), The Pensions Regulator (TPR) and Pension Protection Fund (PPF) issue joint statement on defined benefit pension scheme data

This is a joint statement, published on behalf of the Office for National Statistics (ONS), The Pensions Regulator (TPR), and the Pension Protection Fund (PPF).

While defined benefit (DB) pension scheme data are collected for different purposes, we recognise the importance of broad alignment across the set of statistics which describe the UK DB pension landscape. Compared to that of the ONS, TPR and the PPF have a different dataset and methodology for estimating DB pension scheme funding positions.

The ONS, TPR, and the PPF have worked alongside each other for many years and with increasing closeness since autumn 2022, to understand the differences between our respective datasets on DB pension schemes. We have worked collaboratively to provide this combined response and will continue working together on an ongoing basis.

It is important to acknowledge the differences in the approaches taken to collecting data and the different purposes for which they are used. However, we expect our datasets will show greater alignment in future based on our collaboration and proposed changes.

The ONS uses the Financial Survey of Pension Schemes (FSPS) to collect a sample of occupational pension schemes’ data on a quarterly basis including membership, income and expenditure, assets and liabilities. The data is collected primarily for inclusion in the UK National Accounts and UK Balance of Payments. It also informs the estimate for UK Gross Domestic Product (GDP).

TPR is the public body that protects workplace pensions in the UK. It collects data on all funded occupational pension schemes through two main sources: the recovery plan submission process and scheme return. Under either of these processes, the latest asset and liabilities values are only provided once every three years in accordance with the valuation cycle, and trustees have fifteen months to submit valuation results. As such, on average, the asset and liabilities values held by TPR for the DB pensions universe are between two and three years old.

The PPF is a public corporation run by an independent board and is responsible for providing compensation to members if the employer funding their pension scheme becomes insolvent and the scheme has a shortfall between its assets and liabilities. PPF compensation is generally lower than full scheme benefits because members below retirement age receive 90% of their pension, and there are limits on indexation provided by the PPF. The PPF receives its data from TPR and so has the same assets and liabilities data as TPR which, as above, is on average between two and three years old.

Both TPR and the PPF use their own actuarial modelling techniques to calculate current estimates of the assets and liabilities of DB pension schemes to monitor the assets, liabilities, and funding levels, and judge the associated risk of these schemes.

The following summarises the main differences in data collection and estimates between the three organisations:

1. The ONS FSPS uses a sample of funded occupational pension schemes to collect pension assets data at current dates. As such, the ONS FSPS does not have full coverage of the DB scheme universe, and to estimate the overall DB universe, uses this data to make estimates for the remaining schemes. The ONS FSPS sample has coverage of broadly 80% of the DB universe by asset size.

2. By contrast TPR and the PPF have data for the whole universe. However, given the triennial valuation cycle of DB schemes, at any given time a substantial amount of this data will be up to three years’ old. Both TPR and the PPF instead use their own actuarial modelling to calculate their estimates of current assets and liabilities, and in doing so, adjust both the assets and liabilities to account for changes in financial markets, through standard market indices. Further adjustments are made to their estimated asset values to allow for deficit reduction contributions paid.

3. As the ONS FSPS collects quarterly data, ONS assets data are more up to date in aggregate, reflecting the benefits accrued to and paid by schemes, actual investment returns, changes in asset allocations, and risk transfers between pension schemes and insurance corporations. By contrast, as TPR and PPF data is historical, allowance for the impact on assets of changes in financial markets (investment returns) is estimated using financial indices.

4. The ONS FSPS does not currently publish pension entitlement liabilities data and therefore does not publish estimates of scheme funding levels.

5. By contrast, both TPR and the PPF are interested in the funding level of schemes (the difference between the value of assets and liabilities) as this impacts the level of risk to scheme members and the PPF.

6. TPR is primarily concerned with the funding level of technical provisions (Part 3), for which its regulatory powers are in force. The PPF is concerned with the funding level on a section 179 (PPF) basis, which compares different benefit provision to full scheme benefits.

7. As funding levels look at both liabilities and assets, historically both TPR and the PPF have not made any allowance for either benefits paid to members or the accrual of benefits during the period between the effective date of the valuation and the current date. This is because benefit payments and accrual have historically impacted both assets and liabilities by broadly the same amount and therefore have had a minimal impact on the funding level, which is a key metric when determining risk in the DB pensions universe.

8. TPR does not collect information regarding benefit payments or accrual and requires further estimates or actuarial judgement to make allowances for these factors in their modelling. At the inception of TPR and the PPF, the DB landscape was significantly different. Open DB schemes were common, and it was therefore expected that the difference between benefit payments and accrual would have been relatively small at that time. However, given the closure of many DB schemes in the years since, the latest ONS FSPS estimates show there is a current net outflow of around £30 billion per year from these schemes. Benefit outflows from private sector DB schemes are in the region of £50 billion per year. Contributions (accruals to DB schemes) have been on a downward trend since the beginning of the FSPS timeseries in 2019 and are currently in the region of £20 billion per year.

9. Recent TPR and PPF asset valuations would be expected to be between £80 billion to £120 billion higher than ONS estimates, because of the above factors alone. However, it is important to note that TPR and PPF liabilities valuations would also have been £80 billion to £120 billion too high, such that the overall surplus (or deficit) and funding level would be broadly the same, with or without these adjustments.

10. 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.

While recognising the differences between the datasets and their purposes, we acknowledge the growing interest from stakeholders and the desire for greater alignment across our publications. We appreciate the importance of reaching a shared understanding of the current funding position across the DB pension schemes universe.

Indeed, the House of Commons Work and Pensions Committee recommended that TPR, the PPF and the ONS should continue to work together “to reach an understanding of the funding position of DB schemes and publish the results” in its Third Report of Session 2023 to 2024, published in March 2024.

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. The PPF expects to include this change when it publishes its Purple Book, and then in subsequent PPF 7800 index releases.

This change has been incorporated in response to the increasing focus on comparisons between TPR, the PPF, and ONS data. As TPR does not collect benefit payment or accrual data, TPR has used the data from the ONS FSPS when making these adjustments. The PPF will adopt a similar approach.

Some margin of error should always be expected between the three estimates because of the different modelling approaches taken by each organisation. As such, as set out above, it is important not to compare one set of statistics from one organisation with another’s in isolation to avoid incorrect estimates of funding levels or deficits and surpluses. However, the changes made to TPR and PPF methodologies provide greater alignment between TPR, PPF and ONS estimates of the assets held by DB schemes.

Our analysis suggests that once these changes have been incorporated into TPR and PPF models, ONS, TPR and PPF estimates are close at the dates analysed. We expect the magnitude and direction of the differences in estimates will fluctuate over time.