10 March 2006
Fears of a multibillion pound black hole in public sector pensions have brought forward calls for the chancellor to amend his fiscal rules. John Hawksworth analyses the options and calls on the Treasury to provide more detail on the liabilities and issue a discussion paper
Private sector companies have been forced to face up to their defined benefit occupational pension liabilities in recent years, due in large part to the light shone on them by the introduction of Financial Reporting Standard 17 on retirement benefits. Of course, the public sector has also adopted FRS17 but the amount of readily available information on public service pension liabilities is much more limited.
Moreover, while government departments and other public bodies now use FRS17 in their accounts, Chancellor Gordon Brown's fiscal rules are based on an older set of European national accounting standards used for economic statistics, known as ESA (European system of accounting) 95.
As a result, Brown's 'golden rule' – borrowing only to invest over the course of an economic cycle – requires payments of public service pensions in each year to be covered by current tax revenues (on average over the economic cycle). But it does not make allowance either for the net accrual of public service pension liabilities from year to year, or the implicit cost of funding them.
Similarly, the chancellor's second fiscal rule, to keep net public debt below 40% of national income (gross domestic product), does not take account of public service pension liabilities.
Independent commentators such as Watson Wyatt and Neil Record have put forward estimates of the order of £700bn–£800bn (around 58%-67% of GDP) for already accrued unfunded public service pension liabilities at the end of March 2005. Official estimates published last week suggest the figure might be around £530bn, although this is still up by around £70bn on the previous year.
Focusing on such 'shock horror' numbers is, however, not the best way to encourage a mature debate on this issue, even if it does make for some juicy headlines. These numbers are the cumulative present value of liabilities that will have to be paid over the next 85 years, not money the taxpayer will have to find soon.
More relevant to assessing fiscal sustainability are projections of future public service pension payments as a share of GDP, which the Treasury has been publishing in its annual long-term public finance report for the past two years. As the figure below illustrates, the Treasury estimates that these payments are set to rise from 1.5% of GDP now to around 2.2% by 2035 before levelling off.
Click here for Long-term public spending projections (this will open up a new browser window)
While this 0.7% of GDP rise is not insignificant – it is equivalent to around £8.5bn per annum at today's GDP values – it is relatively modest compared with longer-term increases projected in state pension payments and, even more so, spending on health and long-term care. Here at PricewaterhouseCoopers, our own estimates of the likely rise, like those of the Organisation for Economic Co-operation and Development, are rather higher than those of the Treasury, since the latter does not take account of non-demographic upward pressures on health spending.
Nonetheless, there is a question as to how these longer-term public service pension projections, valuable though they are in putting the scale of the problem in perspective, should be reflected in the fiscal rules that drive tax and spending decisions in the short to medium term.
That is why we thought it would be useful to publish a discussion paper on this issue earlier this year, which has since been discussed with experts in the Treasury and elsewhere with a view to promoting such a debate. Officials in the Treasury have clearly been thinking deeply about these issues for some years. Publishing more information might now be a sensible way forward.
Two other developments will help to push this forward. First, the intention is to publish Whole of Government Accounts for the financial years from 2006/07 onwards, which will include aggregate public service pension liabilities calculated on an FRS17 basis.
Some commentators have already suggested that the government's 40% debt rule should be extended to include public service pension liabilities and these calls might well become louder after the first WGAs are published.
Second, and perhaps less widely known, is the fact that there are ongoing discussions on these issues between international economic statisticians from the International Monetary Fund, the OECD, Eurostat, the United Nations and national statistical offices, which have been published in a series of papers on the UN website over the past two years.
These papers make it clear that an FRS17-type standard for unfunded public sector occupational pensions is likely to be included in the next version of the internationally agreed System of National Accounts on which the UK national accounts – and so the chancellor's golden rule – are based. Any such changes are still several years away, given the need for international agreement on these issues, but it is worth starting to look now at what changes in these national accounting changes mean in practice.
For the golden rule, they would mean that you need to set taxes (on average over each economic cycle) at a level that would cover net accruals of public service pension liabilities at the time of accrual, rather than the time of payment.
You would also need to allow for the implicit cost of funding these liabilities. The economic rationale for this is that accrued pension liabilities are a form of deferred pay in return for consumption of labour resources today, and the golden rule states that current taxes should cover current public consumption for the sake of fairness between generations.
As it happens, this change would make the golden rule significantly harder to hit at present, potentially increasing the current budget deficit by around £10bn or more, although the precise amount would depend on the actuarial assumptions made. Particularly important would be the assumption made on the discount rate used to calculate the present value of public service pension liabilities.
Until last year, this was set by the Government Actuary's Department at 3.5% in real terms, but the rate used has now been reduced to 2.8%, which will increase the present value of pension liabilities, but reduce their implicit annual funding cost.
Some commentators have suggested that the discount rate be reduced much further to match the current very low level of index-linked gilt yields (around 1.2% at the time of writing for 15-20 year maturities), but the economic rationale for this is not clear.
For a pay-as-you-go public pension scheme, the liabilities will be met out of future tax revenues that will tend, in the long run, to grow with the trend of GDP, not with the real gilt yield.
So an assumption of a trend in long-term growth of GDP of perhaps 2%–2.5% would seem the most economically rational real discount rate to use and would avoid changing this assumption every time market gilt yields shift.
However, putting the golden rule on an FRS17-type basis might open up wider issues. In particular, at the moment, the full cash value of public debt interest payments are scored as current spending, including an element just to cover inflation. But if government assets also increase over time with inflation, as you would broadly expect, then this suggests that government net worth is rising over time, which, in turn, implies a potential transfer from current to future generations.
Furthermore, the higher the rate of inflation, the faster that net worth rises under the current convention. It is not at all clear why this should be desirable. Taking the inflation element out of debt interest payments would reduce current spending by around £10bn and reduce the potential impact of FRS17-type accounting for the accrual of public service pension liabilities. Given the large sums involved, this issue merits further attention.
Similar complications arise in relation to the chancellor's second fiscal rule on net public debt. As noted above, some commentators have argued that this should be extended to include public service pension liabilities, although this implies setting a much higher limit.
It is not clear what level would be appropriate, given that international comparators are not readily available for public service pension liabilities, in contrast to public debt levels. More fundamentally, however, it is not clear that this would be appropriate from a macroeconomic perspective.
Imagine, for example, that the government were permanently to increase public sector salaries, and fund this through increased public borrowing. This would be likely to have short-term macroeconomic consequences, because public sector staff would probably spend a significant proportion of this extra salary, which would add to demand pressures in the economy.
This, in turn, would be likely to cause the Monetary Policy Committee to set interest rates slightly higher than would otherwise be the case and gilt yields might also be slightly higher, due to increased public borrowing. This general upward movement in relative UK interest rates would also be likely, other things being equal, to attract short-term capital inflows into sterling assets, so tending to push up the exchange rate. Because the Treasury wants to reduce the risk of destabilising movements in interest and exchange rates, in addition to any concerns about longer-term fiscal sustainability, it is therefore important to avoid significant rises in public debt levels.
In contrast, if the government were, instead, to reward its employees by making their pension accruals more generous, this would be unlikely to lead to the same macroeconomic responses. Public sector workers would be unlikely to spend these gains and an interest rate response from the MPC and the financial markets seems much less likely.
So, lumping together public debt and pension liabilities in a single fiscal rule might muddy the signal in terms of the impact on macroeconomic stability, as well as opening up lots of scope for arguments about the actuarial assumptions used to calculate the present value of pension liabilities that could undermine the credibility of the sustainable investment rule.
This is not to say that more information should not be published on public service pension liabilities – this would be highly desirable – but there is not a strong case yet for changing the current 40% debt rule.
The Treasury will no doubt want to consider these issues carefully, including consultation with public finance specialists outside government, before making any changes to its fiscal rules. A good starting point might be to publish a discussion paper alongside the Budget, setting out and assessing the options, as well as publishing more information in the Budget Red Book on alternative measures of current spending and the current budget balance, using an FRS17-type treatment for public service pensions.
In any event, change is on the way, so it might pay for the Treasury to be ahead of the game here.
John Hawksworth is head of macroeconomics at PricewaterhouseCoopers and the author of a recent discussion paper on this issue
PFmar2006
