According to The Times, the cost of pensions tax relief rose by £5.3bn “last year”, taking the total past £50bn for the first time. The paper’s 7 October report quotes “a source close to the Treasury” saying: “The cost of tax relief is only going in one direction at the moment. There is widespread recognition that it needs to be reversed.”

The figures reported are not new. (We blogged about them in February, when HMRC first published the numbers.) And they are for 2015-16, not “last year”. That isn’t just pedantry – it undermines the argument that the cost “is only going in one direction” and that something must therefore be done. Things that have recently been done were not implemented in time to affect the 2015-16 numbers:

  • The Lifetime Allowance was cut from £1.25 million to £1 million with effect from 2016-17. 
  • At the same time, the Annual Allowance was cut for the highest earners – it now tapers down from £40,000 to £10,000 as income (including employer pension contributions) rises from £150,000 to £210,000.
  • The Money Purchase Annual Allowance for people who have accessed pensions flexibly is being cut from £10,000 to £4,000 with effect from 2017-18, and should apply more widely as the number of people still working who have dipped into their pension pots creeps up. 

It is true that phasing in higher default contributions under automatic enrolment will make more employee compensation tax-deferred and increase the upfront cost of tax relief (as anticipated when the policy was devised). The maturing of closed defined benefit schemes in the private sector works the other way. The projected growth in, and ageing of, the working age population should increase the cost of tax relief – but not without increasing tax revenue itself.

If the claim that things are only heading in once direction is debatable, the “more than £50bn” (meaning £54bn) total is true only on a peculiar construction of the facts. This number is derived by summing four estimates, which HMRC says have “a particularly wide margin of error”, namely:

  • The cost of upfront tax relief on individuals’ pension contributions; plus
  • The cost of not taxing individuals upfront on employer contributions (including contributions to repair deficits in defined benefit schemes); plus
  • The cost of not taxing individuals on investment income within pension funds; plus
  • The cost of not levying employers’ National Insurance Contributions (NICs) or employees’ NICs on employer pension contributions.

In other words, the “cost” of tax and National Insurance relief in the Times report is measured against an imaginary alternative where money saved in pensions comes out of income that has already been subject to tax and NICs, where the investment income is taxed, and where withdrawals are taxed too. Such a system would heavily penalise people for saving money in a pension as opposed to spending it or saving it in another way. In 2014, the Institute for Fiscal Studies said that even where the official number is presented with tax on current pension payments netted off (which would be £40.5bn on the latest numbers), a “better estimate” would be “less than half” as big.

There are reasons why the Chancellor might favour the “pension raid” that the Times says is on the horizon, and reasons why he might not. If there is a change to be justified, HM Treasury might quote figures such as these – you use the ammunition you have. But we hope that premature conclusions about a strangely-chosen number’s trend won’t influence policy decisions.