Public Service Pensions Bill

Memorandum submitted by John Ralfe (PSP 21)

Public Service Pensions Bill


I would like to submit comments on the Public Service Pensions Bill to the Public Bill Committee.


I am an independent pension consultant and have written on many aspects of public sector pensions, in the Financial Times and elsewhere, as well as being interviewed several times on the Today Programme and Channel 4 News. Please see my attached biography.


In making changes to public sector pensions it is crucial that the real annual economic costs of the pension promises are used. I believe that the official annual costs, calculated by the Government Actuary’s Department ("GAD"), and quoted by the government, are materially understated. This means:

a. Proper comparisons between the old and the new public sector pension benefits, such as the impact of increasing the retirement age or increasing the annual accrual rate, are impossible.

b. Proper comparisons between the generosity of public sector pensions and private sector DC pensions, are impossible.

c. At the macro-level, the current generation of taxpayers will pass an economic cost to be paid by future generations, which is inherently unfair.

All of this challenges the idea that the public sector pension changes are fair and equitable with the private sector and represent a "settlement" for 25 years.


The government quotes that its changes to NHS, TPS and Civil Service pensions have reduced the overall cost to taxpayers by a third from around 23% of salary to 15% of salary, which have been repeated in the recent report published by the Pensions Policy Institute. [1]

The PPI estimates that the average private sector DC cost, including contracting into SP2, is 10% of salary, so it concludes that the new public sector pension terms are only 5% more generous than the majority of private sector pensions.


However, I believe these costings are wrong:

a. The total real cost of public sector pensions, before member contributions, remains at around 31% of salary, even after the changes. The saving from the higher retirement age has been offset by the higher rate of pension earned each year.

b. The cost saving for taxpayers is due to the increase in member contributions of around 3%. The increase from just over 6% to just over 9%, reduces the real cost to taxpayers from around 25% of salary to 22%, a much smaller saving than suggested by the government.

c. This in turn means the new public sector pensions at a cost of 22% of salary, versus 10% in private sector DC, are much more generous than suggested by the PPI - 12%, not 5%, more generous.


The GAD calculates annual public sector pension costs using the Treasury method of discounting expected pensions at CPI + 3%, representing expected GDP growth.

The correct discount rate should be based on the yield on long-dated index-linked gilts, since public sector pensions and ILGs share similar characteristics. Both are obligations of the UK government, both are contractually committed, legally-binding and both are inflation-linked. (The yield should be adjusted for the differential between CPI and RPI as public sector pensions are uprated in line with CPI).


I attach a letter to George Osborne in April 2011, signed by 23 pension experts from the UK, US and Australia, arguing for the use of the ILG yield, not forecast GDP growth. This received extensive press coverage, including:

a The Financial Times April 27 th 2011

"Rethink urged on future pension bill" By Norma Cohen

b The Financial Times May 2nd 2011

"Bean counters ignored over discount rates" By Pauline Skypala

c Robert Peston BBC blog May 3rd 2011

"Is the Treasury understating pension liabilities?"


The ILG costings I quote above are based on a real ILG yield of 1%, significantly higher than the current yield, so there is an argument that even these costs are understated.


Using ILG yields to calculate annual pension costs and liabilities is not an academic exercise – it is precisely the method used to value the liabilities of the House of Commons Members' Fund which makes payments to former MPs and dependents with little or no pension under previous arrangements.

The latest valuation by the Government Actuary in September 2006 uses "a market-related approach such that the interest rate used to discount the liabilities falling due in future years is the real yield available in the open market, on the reporting date, on investment in a medium-dated index-linked gilt portfolio. Accordingly, a discount rate of 1.5% a year net of price inflation has been used to value the Fund's liabilities" (para 6.3) [1] . The rate in 2003 was 1.86% reflecting higher gilt yields.


Furthermore, the Bank of England uses ILG rates to calculate its annual pension costs, which were as 54.5% of salaries in the 2008 valuation.


The official total cost of public sector pensions, before member contributions, bears no relationship to the market cost of individual or bulk annuities.


There are many examples of economists and pension experts who support using the riskless rate, or ILGs, in addition to the 23 who sent the letter to George Osborne.

Let me quote Donald Kohn, a Member of the Bank of England’s Financial Policy Committee, who said in 2008, when he was Vice-Chairman of the Federal Reserve Board:

"... public pension benefits are essentially bullet-proof promises to pay. We all have read about instances in which benefits were lost when a private-sector pension sponsor declared bankruptcy and terminated the plan. In the public sector, that just hasn't happened, even when the plan sponsor has run into serious financial difficulty. For all intents and purposes, accrued benefits have turned out to be riskless obligations. While economists are famous for disagreeing with each other on virtually every other conceivable issue, when it comes to this one there is no professional disagreement: The only appropriate way to calculate the present value of a very-low-risk liability is to use a very-low-risk discount rate." [1] (My emphasis. NB although he is talking about the US, the same conclusion also applies to the UK).

David Wilcox, a senior economist at the Washington Federal Reserve, also said in 2008:

"The economics of how cash flows with no credit risk should be discounted are utterly unambiguous and non-controversial. They should be discounted using rates derived from securities with no credit risk." [2]

I would be very happy to discuss these crucial issues with the Committee.

November 2012

26th April 2011

Dear Mr Osborne

Public sector pensions discount rate

You announced in the Budget that the annual cost of new public sector pension promises would be calculated using a discount rate of expected GDP growth above inflation and the formal reasons for this were published on April 6th.

We are writing to ask that you re-consider this decision which we believe fundamentally misrepresents the economics of public sector pensions and has serious pernicious consequences.

In our view the correct discount rate should be based on the yield on long-dated index-linked gilts, (adjusted for the difference between consumer price inflation and retail price inflation), since public sector pensions and index-linked gilts  share similar characteristics. Both are obligations of the UK government, both are contractually committed, legally-binding and both are inflation-linked.

The Consultation suggests the argument for using expected GDP growth is that pensions are "paid for out of future tax revenues ".

But gilt interest and principal payments are also paid for out of future tax revenues. This clearly does not mean that new gilt issues should be valued by discounting payments in line with expected GDP growth, rather than the market gilt rate.

In using expected GDP growth, the Treasury has not explained how an obligation to pay a public sector pension differs from an obligation to pay gilts. If there is no difference, then pensions should be discounted at the gilt rate. The other possibility, that gilt payments should be discounted at the expected GDP growth rate, is immediately contradicted by the market.

The government’s approach implies that it is cheaper for it to promise an inflation-linked pension payment to a public sector employee than it is to pay the coupon and principal on an index-linked bond.

By overstating the discount rate we understate both the current economic cost of public sector pensions and the real economic savings from the Hutton Report’s recommendations. It also means that the efficiency of individual public sector bodies is overstated, as employment costs are understated and at the macro-level, the current generation of taxpayers is passing on an economic cost to be paid by future generations.

We must be clear that public sector pensions are not discretionary government spending, like health or education, which, subject to the ballot box, can be reduced to maintain affordability. They are deferred pay earned as part of a legally binding contract of employment, the equivalent of giving gilts to be redeemed at retirement and we believe their true cost should be properly measured.

In light of this we ask you to re-consider this decision.





[2] Comments before the Public Interest Committee of the American Academy of Actuaries September 4, 2008



Prepared 21st November 2012