Iain Duncan Smith’s cat

In the wake of yet more delays and questions about value for money, I wonder what Erwin Schrödinger would have made of universal credit.

In the Austrian physicist’s famous thought experiment, a cat is placed in a sealed box with a flask of poison and a radioactive substance. The decay of a single atom of the substance during the test will trigger a hammer that breaks the flask and kills the cat. The point is that an external observer cannot know whether or not the atom has decayed, the poison has been released and the cat is dead unless they open the box. Since we cannot know, the cat is both alive and dead.

Schrödinger’s Cat was meant to illustrate a paradox in quantum theory but it could just as easily be applied to Iain Duncan Smith’s flagship welfare reform. It’s not just that universal credit is meant to be simple and transparent but is actually fiendishly complicated and impossible for outsiders to understand. These have become givens over the last couple of years. IDS’s cat also exists in two states at the same time and we cannot know whether it is alive or dead until we open the box or see it in action.

So universal credit is a triumph. IDS told us so in interviews and articles yesterday about ‘the end of the dole as we know it’. The delays and problems so far show that the DWP is willing to learn from experience. The ‘roll-out’ was being accelerated to families that very day and Duncan Smith told the Today programme that he will ‘land’ it (though it wasn’t clear how you can land something that hasn’t taken off yet). The problems are in the past and the critics are all being too negative and universal credit is value for money.

The announcement was a blatant piece of pre-emptive PR ahead of a progress report due today from the National Audit Office. This concludes that: ‘In principle, the Department’s approach should allow it to learn from experience, improve the design and readiness of services and reduce risks. However, in our view the Programme is at too early a stage to determine if the Department will achieve value for money in its implementation of Universal Credit. ‘

It seems the universal credit is a not-triumph at the same time as it is a triumph. The report also reveals further delays to the roll-out. This now extends to the end of 2019 including a two-year delay in the timetable for incorporating tax credits that should reduce risks. However, the DWP does not yet have any plans for when 550,000 of the existing claimants of tax credits and employment and support allowance will move over. This graph from the report shows the declining universal credit caseload expected in 2012, 2013 and now:

In line with Schrödinger’s experiment, universal credit also exists simultaneously in two different forms. The ‘live service’ one was introduced in the North West for single people, then couples, including some with housing benefit claims, and from yesterday families with children. It is being ‘rolled out’ nationwide to simple cases in 2015 and 2016. However, all this relies on old IT systems from before the 2013 ‘reset’. This will result in administrative costs of £149 million but the DWP believes it will generate societal and distributional benefits of £267 million.

The DWP says the nationwide roll-out will give it a chance to form working relationships with local authorities and housing associations. Before then, it is working on changes to the social security regulations about sharing information with landlords, admitting that ‘some landlords and claimants have struggled with rent arrears where support for housing costs is including in a single payment direct to claimants’.

The ‘digital’ version of universal credit – the way it is meant to work – is still barely off the drawing board and the NAO report reveals:

‘The Department’s digital service has been delayed and is still in the very early stages of development but is soon to be tested with all claimant types, even the most complex. Recruitment and capacity problems have delayed the new digital service by six months compared with plans at the start of 2014, and it has not yet reached its planned staffing level.’

Testing of aspects of the digital service starts this month but it can still only handle a small number of claims and it depends heavily on manual intervention. Testing at scale has slipped to May 2016 but the NAO describes the 18-month timetable to move from a few hundred claimants to 10 million as ‘challenging’.

Given the record of universal credit so far, none of this can be taken for granted. A delay of a further six months would lose £2.3 billion in societal benefits and cost £2.8 billion more in staff time for using live service rather than digital systems. However, the NAO comments: ‘The Department does not yet have a plan should the digital service fail and has not evaluated whether it could use live service systems instead.’

This twin-track approach costs more and most of the IT developed for the live service version will not be useable in the digital one. However, the NAO says the DWP ‘estimates that the twin-track approach yields a higher net present value overall by bringing forward the benefits of the programme’. The DWP says they are ‘complementary parts of an integrated approach’ but the NAO says the twin-track programme ‘creates uncertainty about the relative emphasis on the digital service as the end state for universal credit, or a more traditional view of benefit reform with planned updates to IT systems’.

These £20.7 billion benefits to the economy are at the heart of the outline business case for the project that the Treasury finally signed off in September.  They include higher earnings for people as they move into work and reduced spending on benefits. However, the NAO warns that the estimate is ‘heavily dependent on a number of assumptions’ and that the uncertainty is ‘magnified’ by the use of distributional impacts (people on lower incomes are assumed to value a change more than those on higher incomes).

In another duality, universal credit was designed from the beginning to deal with people in work and out of work. A key part of its appeal was that it would simplify the transition between the two and reduce the benefits trap when people take a job. That’s even more vital at a time when more and more people who are working are also in poverty. The DWP expects the majority of people on universal credit to be in work. However, there is also a stick to that carrot: the expansion of conditionality to people who are in work but not deemed to be working hard enough.

That makes it remarkable that the NAO report reveals that the business case does not currently include funding to support working claimants:

‘The Department has not confirmed plans for how it will support claimants who are in work, even though it expects that around one million in-work claimants will have conditions attached to their claim.’

The overall verdict of the NAO is that the DWP has reduced delivery risks since last year’s ‘reset’ by extending the roll-out and choosing the more expensive twin-track approach to developing the service. However, it concludes:

‘We consider it important that the Department, having reset the programme on a sounder basis at significant costs in terms of resource and elapsed time, confirms its plans for delivering Universal Credit in terms of cost, time and functionality, against which it can be held to account for the good use of public resources.’

In other words, IDS’s cat is still in its box, alive and dead, failing and succeeding at the same time. The universal credit paradox is that we cannot know for certain which state it is in and whether it’s gone too far to be allowed to fail or cut our losses and abandon it. The time to open the box could just possibly come some time in late May 2015.