If Universal Credit were made into a Hollywood rom-com, you just know that Margaret Hodge and Robert Devereux would be the central characters.
Thrust into repeated opposition on different sides of a heated Public Accounts Committee (PAC) table, the two star-crossed combatants would bicker and argue in ever-increasing circles of conflict. Then, in the final act, in a moment of outrageous serendipity – such as Universal Credit actually going fully live – they would realise their true feelings for each other.
You can decide which outcome – Universal Credit going live, or a future romance between PAC chair Margaret Hodge, MP and Department for Work and Pensions (DWP) permanent secretary Robert Devereux – is the more likely. (Remember this – nobody would have believed John Major and Edwina Currie…)
Last week the soap opera continued as Devereux was once more hauled in front of MPs by Hodge to discuss the latest National Audit Office (NAO) report into the troubled welfare reform programme. The latest act ended with Hodge insisting Universal Credit is a shambles, and then shutting down the meeting before Devereux had one last opportunity to deny the accusation. For the two pugilists, “shambles” is a binary measure – it either is (Hodge) or it absolutely isn’t (Devereux).
On Universal Credit at least, it’s increasingly clear that “shambolic” is in fact a sliding scale, with the programme veering in one direction or the other to different degrees of shambles. Let’s call it the Devereux-Hodge Scale of Shambolicness, to represent the end-points at either extreme, where 100% Devereux represents everything tickety-boo and working as expected; and 100% Hodge representing a total shambles.
So, based on recent revelations from the NAO and the PAC hearing, where on the scale is the project right now?
Wasted spending or value for money?
The discussions last week in the PAC meeting between Devereux, the Treasury’s Sharon White, the NAO and Hodge veered into the arcane terminology of accounting and economic modelling. At one stage, Hodge highlighted that £697m has been spent on Universal Credit so far – that’s all costs, not just IT costs – and yet DWP said only £34m of that will make it onto the balance sheet as an asset as a result of the “twin-track” approach now being taken, and once the digital system currently being developed is live.
On face value, that certainly seems like close to 100% Hodge on the shambolicness scale. Some reports claimed this means DWP could “write off” £663m, but that’s not the case. As White – soon to be leaving the Treasury to become the new CEO of Ofcom – explained, much of that £697m is normal operating expenditure which would never be recorded as an asset, and part of it is “Plan B” contingency spending so the existing Pathfinder system being used can stay in place if the digital system is delayed or doesn’t work.
We know for sure that £131m of assets will be written off by the time Universal Credit is fully live, but it is certainly likely that the true figure will be much higher.
Much of that £697m has gone on things like staff costs – money which would have been spent anyway – and up-front design and planning work. The only way that expenditure will have been wasted is if Universal Credit is scrapped entirely. That’s unlikely as Labour is entirely supportive of the policy, if not the implementation plan.
But Hodge is absolutely right to say that spending £697m and still not yet having a fully agreed business case is a shambolic way to spend money. In the unique language of the Civil Service, DWP has so far had its Strategic Outline Business Case approved; next year the Outline Business Case is due for approval; and then in 2016, the Business Case should finally be approved.
Devereux, on the other hand, maintains that spending £697m is chicken feed compared to the £7.7bn that will be saved as a result of the twin-track approach compared to the other alternatives considered when the programme was “reset” last year. That £7.7bn figure is derived from those arcane economic models based on more benefit claimants being on Universal Credit sooner, and the assumption that more of them will go back into employment more quickly.
If you want to consider the programme on the basis of a good old-fashioned return on investment calculation, then Devereux’s figures make such spending seem minor compared to the promised returns. This has, consistently, been the DWP line – that for all the spending, all the write-offs, and all the delays, the benefits to the UK of Universal Credit vastly outweigh the problems in getting us there.
So how do you decide where this stands on the Devereux-Hodge scale? You could consider the views of the independent Office for Budgetary Responsibility (OBR), which stated last week that in its opinion, “there remains considerable uncertainty” around the plans for Universal Credit, and that “weighed against the recent history of optimism bias in Universal Credit” it expects at least a further six-month delay in the latest revised timescales – which themselves are considerably delayed compared to the original plans back when the project was launched in 2011.
So in terms of the value for money so far, we’re very much at the Hodge end of the shambles scale – but if you’re looking long term (and believe the DWP economists) then it’s closer to the Devereux end.
The business case; or, does anyone actually know what they are doing?
As mentioned above, the phrase “business case” seems to have a flexible definition in the Civil Service. But there is no disputing that DWP has yet to produce what the NAO calls a “target operating model” for Universal Credit.
In layman’s English, this effectively means the DWP has yet to define what Universal Credit will actually do, how it will work, and what its processes and workflows will be. And you’d have to say that if this were the business case in a company, you wouldn’t get the project past first base if you had not defined what the end result would be. Even an agile project has a good idea of the desired outcome.
One of the independent advisors to Labour’s review of Universal Credit told me that for any project of this scale, failing to determine right up front what the target operating model will be and how the future processes will work, is a fundamental error. Labour has promised a three-month pause to the project should it win the general election in 2015, and one of the primary reasons for that is to step back and define that target model.
DWP would counter that it is pursuing a “test and learn” approach – a vaguely agile concept whereby it introduces new features and functions, then learns from their trial implementation, feeding the results back into the roll-out process. Bear in mind of course, that “test and learn” only came about when the project was so out of control that it was halted, reviewed and “reset” because it was without any doubt 100% a Hodge-level shambles.
In many ways, this comes down to the IT argument of waterfall versus agile, and if ever a project was trying to shoehorn both approaches into one, it’s Universal Credit. Agile is not a panacea; as the Labour advisor put it, it’s “horses for courses”. And in its desire to be seen to be agile, Universal Credit forgot some of the basics – namely, agreeing up front what they were all meant to be aiming for.
The NAO’s previous September 2013 report made the same criticism, stating: “Throughout the programme the department has lacked a detailed view of how Universal Credit is meant to work… The department was warned repeatedly about the lack of a detailed ‘blueprint’, ‘architecture’ or ‘target operating model’ for Universal Credit.”
Work has started to address that gap – the Treasury refused to sign off even the Strategic Outline Business Case without it – but the truth is that no matter how much testing and learning takes place, it’s difficult to say with certainty how much of the work completed so far will be relevant for a target operating model that has yet to be fully defined.
So on the business case, you’re looking at 80% Hodge so far.
The digital Holy Grail
The digital system being developed as part of the twin-track approach will eventually replace all but £34m of the IT assets created to support the Pathfinder trials – that’s just 17% of the £196m of IT assets created so far being retained, according to the NAO.
Digital has become the Holy Grail for Universal Credit – the knight in shining armour riding over the horizon to rescue the programme. It’s easy for this to be the bright future when it’s only just passed its “alpha” stage of development, and the initial trial of the system in Sutton is processing just 17 claimants so far, with a fair amount of manual intervention still required. It is simply too soon to tell how well the digital development is going – especially since the pilot started six months later than planned, based on a timescale established only 12 months ago.
Insiders are saying good things about the management of the digital project under DWP’s digital transformation director Kevin Cunnington. The reason for the delay came from problems recruiting suitably skilled digital expertise into DWP – Computer Weekly reported back in January that the initial recruitment plans were already proving over-optimistic.
It’s worth remembering at this point, that the Cabinet Office and the Government Digital Service (GDS) had recommended that DWP put all its Universal Credit eggs into the digital basket, and scrap the Pathfinder system entirely. The latest NAO report shows that the DWP has managed to play with its accounting and economic models well enough to demonstrate that the twin-track approach will save the UK £7.7bn more than if they had waited for the digital system to be ready and stopped the current system roll-out last year, as GDS advised.
At the PAC meeting, Hodge set out her wariness over the optimistic promises around the digital system, compared with the reality of how the programme has gone in the past. For Devereux, digital can be the bright future for only a short time before it has to prove itself.
The NAO, meanwhile, warned that failure to complete the digital system, and relying instead on the existing system for full roll-out, comes with a £2.8bn bill to taxpayers.
So, 50-50 on the Devereux-Hodge scale so far.
The moving target of roll-out timescales
Gauging the shambolicness of the Universal Credit roll-out to date depends very much on the tint of the glasses through which you view progress.
Based on the original plans, over four million claimants were meant to be on Universal Credit by April 2014. So far, fewer than 18,000 people are claiming the benefit. By any standard, that is 100% Hodge of a shambles.
But through DWP-tinted glasses, it is far more important to get Universal Credit right in the end, than to adhere to an unachievable timescale – even if that was DWP’s own timescale you’re talking about.
As with the digital system, it’s easy for DWP to be confident about its latest roll-out plans because most of the work (and the risk) is so far away that within the confines of a House of Commons committee room everything can easily be 100% Devereux. Even Hodge finds it hard to disagree when asked, “What would you prefer, that we get it right in the end, or we get it wrong on any timescale?”
The roll-out plans are still very risky because they are so back-ended. Millions of claimants will have to be migrated onto new systems over an 18-24 month period – and even then, migration of tax credits has been further delayed until 2019. That scale of migration is simply unprecedented under any government.
The NAO report said that a further six-month delay (which is what the OBR expects to happen) will mean the loss of £2.3bn in potential economic benefits to the UK.
If nothing else, you have to give DWP credit for not blindly sticking to its wildly unrealistic former timescales. But based on progress to date, it’s been a comfortable 75% Hodge shambles – and the risk of further delays is at the same end of the scale.
The Devereux-Hodge Shambles Rating
Overall, based on progress to date, Universal Credit sits very much nearer to the Hodge extreme of shambles than to Devereux. The DWP’s argument is based very heavily on future promises, of jam tomorrow – and of course, if they are right, that’s going to be one heck of a tasty jam sandwich.
It’s not too late to swing Universal Credit towards the Devereux end of the scale. But with national roll-out of the most basic benefit claims due to take place in 2015, and the early promise of the digital system set for its biggest test, we will learn in the next six to 12 months if the needle is going to move along the scale towards Devereux any time soon.