Is the Treasury understating pension liabilities?
Belatedly, I've got round to looking at the Treasury's recent decision to change how it calculates the necessary contributions that have to be made to cover the future costs of unfunded public service pensions.
My interest was sparked by a letter sent to the chancellor by 23 pension experts, organised by the consultant John Ralfe. They argue that the Treasury has made a mistake in its choice of a new so-called discount rate.
If you think this is tedious abstruse stuff that has no relevance to you, think again. The aggregate public-sector net liability for pensions is so huge - perhaps £1 trillion - that it matters to all of us as taxpayers, especially those likely to be paying tax in 10 and 20 years time, that the government has a reliable and accurate valuation of pension promises.
Pensions represent, to coin the phrase, a massive off-balance-sheet debt. And as we've all learned to our cost from the financial crisis of 2007-8, it is a bad idea to carry on blithely pretending off-balance-sheet liabilities don't exist.
So what is this blessed discount rate? Well in the private sector it can be seen as the number used to translate into today's money a commitment to pay £650 a week pension (for example) for 30 years or so to a retired employee (till he or she dies), so that we can see whether there's enough money in the pension fund to pay that employee (and all the other employees) during his or her long retirement.
The point of the discount rate is to assess whether there's enough money in the pension fund - or whether it needs to be topped up.
Which is all very well, except that for most of the public sector, there are no funds or pots of money to pay for future pensions. Most of the pension promises are unfunded, payable out of employees' current contributions and out of general taxation.
That said, since public sector workers are increasingly expected to make a contribution to the costs of their own pensions, it would presumably be sensible for that contribution to be set at a level that is rationally related to the value of promised pensions.
So what is the best way of measuring the cost today of new pension promises?
Well the government has decided to "discount" those promises by the rate at which the economy is expected to grow.
Now there is some logic to that: the growth rate of the economy should determine the growth rate of tax revenues; and the growth rate of tax revenues will have a direct bearing on whether future pension promises will bankrupt us all or not.
But here's the thing. Any private sector chief executive might well be sent to prison if he or she decided to use the equivalent discount rate for a company, which would be the expected growth rate of that company's revenues or profits.
The reason is that although it might be possible to remove subjectivity (or in a worst case, manipulation) from any long-term forecast of the growth of GDP or of a company's turnover, it is not possible to remove considerable uncertainty.
To illustrate, the Treasury has chosen a GDP growth rate of 3% per annum as the discount rate for public sector pensions, which is considerably above the rate at which the UK economy has grown for years or indeed may grow for many years.
If we were growing at 3%, we would in practice be less worried about the off-balance-sheet liabilities of public-sector pensions, because the on-balance-sheet debt of the government would not be growing at an unsustainably fast rate.
To put it another way, in choosing its view of the long term growth rate of GDP as the discount rate, the Treasury is arguably understating the burden of future pensions to a considerable extent.
So what discount rate do companies use?
Well they are obliged to discount the liabilities at the yield or interest rate on AA rated corporate bonds.
Which may not be ideal, but has some advantages: there is a market price for AA corporate bonds, so the yield or discount rate is difficult to manipulate by unscrupulous employers; and it tells the company how much money would need to be in the pension pot, on the basis that all the money were invested in relatively safe investments (AA corporate bonds).
Now Ralfe and his chums believe that the discount rate for public sector promises should be the yield on long-term index linked gilts (gilts are bonds or debts of the British government) - partly because this too has a difficult-to-manipulate market price and because an index-linked government bond is a very similar liability to a public sector pension promise (both are protected against inflation, both are in effect debts of the government).
They point out that gilt interest and principal payments are paid out of future tax revenues, just as future pensions are. So if the value today of future pensions should be discounted at the GDP rate, that's how index linked gilts should be value on the government's balance sheet - which would be bonkers.
Anyway, if you've read this far (and many congratulations to you if you have), you may take the view that it would not be rational to impose a tougher discount rate on the government than on private-sector companies - which is what Ralfe et al seem to want, in that the yield on index linked gilts will always be lower than the yield on AA corporate bonds (because HMG, even with all its debts, is deemed to be more creditworthy than any British business).
But for a government and for a chancellor who have made it a badge of honour to bring transparency and prudence to public-sector finances, prospective GDP growth does look a slightly rum discount rate for valuing those enormous pension liabilities.