Universal Credit: there are reasons to question the supposed net benefit of £38bn

I’m grateful to Andrew Jones for drawing my attention to a table in the NAO report (Figure 2, page 15) which explains how the DWP proposes to find a net benefit of £38bn from the current  annual expenditure of £65bn on the benefits that will become Universal Credit.  The figures look like this:

Net benefits of Universal Credit (£ billion, 2011-12 prices)

Twelve years from 2010-11 to 2022-23 Annual impact from 2022-23
Total saving (cost) to government (DEL) (0.6) 0.4
Total saving (cost) to government (AME) 10.8 2.3
Total benefits (cost) to wider society 27.8 4.4
Net benefits 38.0 7.1
Net present value 27.0 4.7


The benefits are supposed to be over 12 years, so that the actual benefit is just over £3bn a year, rising to £7bn after the system beds down; more than 60% of the benefits, and nearly three-quarters in the initial period, are to the ‘wider society’.

Quite apart from iffy use of a 12-year period, and the long lead times, there are two other problems with the figures as they stand.  First, the  scheme has already been delayed, and a system that isn’t up and running can’t possibly deliver the benefits claimed.  As the schedule is now at least two years behind, two years of the final benefit disappear.  Second, net present benefit is supposed – as the term implies – to be net.  In the plus column go the benefits to society; in the minus column go the costs, to anyone who experiences them.  That’s how cost-benefit appraisal works.  If claimants are getting less, that is a cost to claimants.  Taking these two points together, the apparent net benefit falls, on the government’s own figures, from the claimed £38bn to £19bn: that is, £27.8 bn minus twice £4.4 billion.

That, of course, is only what’s apparent.  It’s rather difficult to say how good the remaining figures are, because the basis of the DWP’s Dec 2012 ‘economic case’ hasn’t been made public.  What is publicly available is the Dec 2012 Impact Assessment, which includes hardly anything of this, but which states overall (and rather more plausibly) that the measure has a net cost  of zero.  If the impact statement is right, and the government expects to save £0.2bn per annum on fraud and error, there is still another £18bn of supposed benefits to account for.

Giving evidence in September to the Public Accounts Committee, Sharon White of the Major Projects Authority said that the Treasury had not actually seen the business case: “We expect the net benefits to be substantial,  but we shall not sign off a specific number until we see the final business case.”   It seems, then, that no justification has been offered to the Treasury for the claims of social benefit; and that makes it look, bluntly, as if the figures have  been made up.

3 thoughts on “Universal Credit: there are reasons to question the supposed net benefit of £38bn”

  1. Paul,

    Re: If claimants are getting less, that is a cost to claimants..

    I think you have forgotten that the theory is that claimants will only get less benefits if they are earning more through work. When they are earning they will receive a top-up ‘universal credit’ so that they always have more income. If there had been such a big hole in the Cost-Benefit Analysis the NAO would have spotted it. Nice try though!

    Re: The Major Projects Authority said that the Treasury had not actually seen the business case

    Not quite right. They have seen the original business case (as has NAO obviously) – they just haven’t seen any updated version since the project was frozen, That is because DWP has not yet worked out how they are going to continue. DWP has promised new plans and a new business case before Xmas.


    1. Brian’s quite right. Looking at the NAO’s Appendix Four (page 47 of the report linked above) the DWP has put up the supposed benefit from increased employment from £3bn (Oct 2011 estimate) to £18.5bn (Dec 2012 estimate). It looks to me like another fantasy figure, but it does make an enormous difference to the calculation.

      I should also note, from Appendix Four, that the bulk of the £27.8 bn is based on £18.6bn ‘distributional (non-cash) benefits’. This, like several other figures, seems to have gone up and down in the estimates like a yo-yo. Having said that, the most unbelievable figure may just be the most stable – the supposed reduction in overpayments through fraud and error of £1.2bn a year. There’s no process by which this can happen.

Leave a Reply