Ed Humpherson to Liz McKeown: Wealth and Assets Survey – valuation of defined benefit pension wealth

Dear Liz

Wealth and Assets Survey: valuation of defined benefit pension wealth

I am writing following our recent consideration of concerns raised with OSR about the valuation of defined benefit (DB) pension wealth in the Wealth and Assets Survey (WAS), and the further engagement we have undertaken with a range of expert users and stakeholders. 

As you know, OSR considered the DB pensions methodology change for Round 8 data (April 2020 to March 2022) in our June 2025 compliance review of WAS. We recognised that ONS had commissioned independent actuarial advice from the Government Actuary’s Department (GSD) and had accepted the resulting recommendations. We also recognised that methodological improvements of this kind involve trade-offs between accuracy, stability, continuity, practicality and user need. Our review also highlighted the continuing need for clear communication, effective user engagement and comparability over time, including the need to revisit the decision on producing a back series. We note that ONS has now published the back series of data to include the impact of applying the updated pension methodology on pension and total wealth for Round 7 data (April 2018 to March 2020).

Since our review, concerns have continued to be raised with us about the use of a SCAPE-based discount rate to value unfunded defined benefit pension wealth. To inform our thinking, we have undertaken further engagement with a range of stakeholders and experts holding differing views on this issue. 

That engagement highlighted that disagreement is not confined to technical implementation details. Rather, it reflects different views about the purpose of the Wealth and Assets Survey and the concept of wealth which the survey is seeking to measure. 

A number of users, including those whose primary interest is distributional analysis, argued that household wealth should be valued on a basis that is comparable with other assets held by households. On this view, market-based discount rates are important because they preserve comparability across asset types and between households. These users expressed concerns that a growth-based approach may not provide the most appropriate measure of household pension wealth for their analytical purposes. 

By contrast, other experts emphasised that unfunded public-sector defined benefit pension entitlements do not have an observable market price and that no wholly satisfactory market comparator exists. They questioned whether large changes in estimated wealth arising solely from movements in market interest rates provide meaningful information about changes in the underlying economic position of pension holders. They also highlighted the different needs of users interested in changes over time and the potential importance of stability and interpretability for some uses of the data. 

We were also struck by the fact that different stakeholders often approached the issue from different perspectives: some focused primarily on the position of the beneficiary, others on the cost to the sponsor, and others on consistency with wider statistical frameworks. These perspectives do not always lead to the same conclusions. 

The range of views demonstrates that this remains an important area of methodological judgement where experts reach different conclusions. In those circumstances, it is particularly important that ONS is transparent about the choices it has made, the trade-offs involved, the user needs being prioritised and the limitations of the resulting estimates. We therefore ask ONS to take the following actions 

• Facilitate a focused discussion with users and experts to test the range of perspectives on the appropriate valuation approach for DB pension wealth and the uses to which the estimates are put. 
• Build on existing published material and clearly and concisely explain the rationale for the current approach to help users make best use of the data. This should include the conceptual rationale for using the SCAPE-based rate in WAS, how ONS has considered alternative valuation approaches, including market-based valuation, and the relationship between WAS and National Accounts estimates. ONS should consider whether additional supporting material, sensitivity analysis, alternative presentations or other explanatory content would help improve transparency, address areas of continuing user uncertainty, and support user understanding of the effect of different valuation choices.
• Use that engagement and analysis to inform future methodological decisions, setting out how user views and evidence will be considered in the next review of the WAS pensions methodology. 

We welcome the progress ONS has already made in responding to the requirements from our June 2025 review, including work on back series, coherence and stakeholder engagement. The points above are intended to support that work and to help ensure that future decisions on DB pension wealth methodology are transparent, well evidenced and informed by the full range of user needs. 

I would welcome a response setting out how ONS intends to take these points forward.  

Yours sincerely,

Ed Humpherson
Director General for OSR