The future of RPI – First Actuarial responds to Government consultation

16 July 2020

First Actuarial has made a formal response to a Government consultation on possible reform of the Retail Prices Index (RPI). Prompted by a 2019 recommendation from the UK Statistics Authority, the consultation identifies shortcomings in RPI, the UK’s oldest measure of inflation, and proposes to modify it. In our response, we raise a number of pension-related and broader questions about the proposed change.

The proposal recommends bringing RPI into line with the methods and data sources of the CPIH – a measure of UK consumer price inflation which includes owner occupiers’ housing costs.

In recent years, the National Statistician has discouraged the use of RPI, viewing it as “a very poor measure of general inflation”. However, it remains widely used. By their own admission, both Government and the UK Statistics Authority lack knowledge of the full range of uses of RPI in the financial and economic life of the country, introducing the risk of unintended consequences, so a wide consultation makes good sense.

In its response, First Actuarial questions the changes to RPI, draws out potential consequences for pension schemes, and makes a number of broader points about the proposed change:

What do the proposed changes mean for pensions?

One major use of RPI by Government lies in its issuance of index-linked gilts. These are widely used by Defined Benefit pension funds seeking to match a variety of inflation-linked liabilities. So the change can affect a scheme’s assets.

Scheme benefits are often broadly linked to RPI inflation (such as salary growth), set at RPI inflation with or without a limit (such as pension increases), or linked to different measures of inflation, such as CPI. So the proposed change can affect a scheme’s benefits.

It is not possible to consider the overall impact of this change on a ‘typical’ pension scheme. Whether a particular scheme ends up better or worse off from the substitution of CPIH for RPI will depend on the nature of both the assets it holds and the benefits it provides. The sponsoring employer could be made better or worse off, with a positive or negative effect on its business.

If a pension scheme deficit increases as a result, larger immediate cash contributions may be required from employers in the short term. In addition, any deterioration in pension scheme funding may trigger a decision to increase active members’ contributions. It is becoming more common for an increase in the contribution rate to be shared between the employer and active members. This can be intergenerationally unfair.

Although we foresee a varied impact on schemes, we can say with certainty that for any scheme holding index-linked gilts, the effect of a reduction in the inflation index used will result in more funding for Government and less for scheme sponsors and members.

This is particularly significant given that regulation has increasingly encouraged pension schemes to hold (and leverage up) inflation-linked gilts, due to the need to ‘hedge their balance sheets’ from one valuation date to the next. Without that regulatory pressure, pension schemes could well have used other assets that provide more general protection against inflation, such as property, commodities and equities.

In combination, the UK Statistics Authority (UKSA) and regulation have in effect caused pension schemes to buy something from Government that could turn out to be, in hindsight, overpriced. The loss to pension schemes and their members is Government’s gain. The reduction in the value of UK government borrowing is around £100 billion, assuming that the impact is to reduce the rate of growth on the inflation measure used by 1% pa beyond 2030.

Beyond pensions – Three broader questions to be addressed

Our response highlights three more general issues, beyond the world of pensions. We believe that before going ahead with the substitution of CPIH for RPI, these broader questions need to be resolved:

• The measurement of housing cost inflation

  Neither CPI (which takes no account of owner-occupier housing costs except for minor repairs) nor CPIH (which uses rental equivalence and council tax) is generally accepted as offering an effective measurement of housing cost inflation. RPI (which includes mortgage interest costs, depreciation, buildings insurance, ground rent and house purchase costs such as estate agents and conveyancing) appears to offer a more comprehensive approach.

• The collection of price data

  The way in which price data is collected has a material impact on an inflation index, as demonstrated in 2010 when a change to the collection of clothing price data increased the gap between the CPI and RPI. We understand that a large-scale reform of the collection of price data is being developed, moving to the electronic collection of huge volumes of price data, away from the manual method of visiting shops. This may alter the results of both CPI and RPI calculations, narrowing the gap between the two.

• Differing objectives of inflation indices

  RPI and CPI are compiled to meet two different objectives. Whereas RPI is devised as a measure of inflation experienced by households, CPI is a macroeconomic inflation measure. The Royal Statistical Society says that both the CPI and CPIH “are an unsatisfactory measure of inflation as it affects British households”. We believe that an existing project referred to in the consultation document, Household Cost Indices, should be completed before substituting RPI, a measure of household cost inflation, with the CPIH, which has a different objective.

Only once these three issues are addressed can the country move forward with an informed decision on inflation indices.
The proposal would also benefit from more advanced, detailed reasoning into the differences between the RPI and CPI, rather than attributing them to flaws in the RPI.

The consultation document draws on limited resources, and does not mention Measuring inflation, a report prepared by the House of Lords Economic Affairs Committee. The Committee took evidence from many sources and made a number of recommendations. Addressing the Committee’s report in detail would be a positive move at this stage, we believe.

In the meantime, we should maintain RPI, but within a continuing programme of development as opposed to a calculation based on fossilised methods.

A significant merit of RPI for research purposes is its continuity for a long period of time. This is particularly important for long-term research where RPI is the only continuous statistic available to allow investigations of relationships between inflation and other measures. Removing an RPI measure will make such long-term research much harder.

Finally, we do have concerns that the proposal lacks transparency. Any contract referencing RPI is materially altered by the substitution of CPIH for RPI, changing the balance of financial interests between the parties. Pension promises made based on RPI are altered for the worse.

The consultation document refers to the observation of a previous Chancellor that “the Authority’s proposal to address the shortcomings in the RPI in this way ‘may be a more efficient approach then continuing to ask users to stop using it and rewriting existing contracts’”. While this approach may be “efficient”, it is unlikely to increase public trust or confidence in either Government or the UKSA. The consultation hints at the avoidance of public debate on the issue – an unhelpful position for Government to take at a time when our society is in critical need of public trust.

Recommended reading

First Actuarial response to the RPI consultation
A consultation on the reform to Retail Prices Index methodology | HM Treasury
UK consumer price statistics: A review | Paul Johnson, Institute for Fiscal Studies
Measuring inflation | House of Lords Economic Affairs Committee
Developing the Household Cost Indices (HCIs) | Office for National Statistics