Universal Credit failures put coalition ICT strategy in purdah
As self-appointed guardian against government IT bodges, the Cabinet Office played a role in deciding Universal Credit could meet its over-ambitious aims. It was not only wrong. It also shares the blame for bringing Universal Credit to its knees.
While Cabinet Office minister Francis Maude and his senior officials were overseeing plans for Universal Credit in 2011, they were simultaneously trying to force their own IT-led political reforms on the Department for Work and Pensions.
The extra work was too much. These initiatives became the most significant points of failure in Universal Credit.
It is incredible the Cabinet Office allowed Universal Credit to go ahead at all on its over-ambitious terms: £2.4bn and two years until full roll-out.
Universal Credit would not only re-engineer the complex administration of £70bn social security payments to 8m households, merging six benefits systems across two government departments and local authorities throughout the country.
It would also rely on councils up and down the country making their own systems and processes compatible. It would depend upon HM Revenue & Customs completing its own income tax system reforms of unprecedented ambition, Real-Time Information. And HMRC would in turn depend on employers, banks and payroll software suppliers reforming their computer systems and processes as well. Universal Credit would have to navigate this exponentially explosive collection of risk factors. This was a spaghetti junction of high-stakes computer gambits from which Universal Credit would have to pull a benefits payment that was reliable and secure.
And that was just the half of it. Apparently despite its job as guardian against IT bodge, Cabinet Office required DWP to simultaneously implement its Digital by Default strategy – a budget-cutting scheme to replace comparatively expensive call-centre and walk-in government services with self-service web apps. This was ambitious enough on its own: root and branch process re-engineering and painful organisational culling of people and premises.
On top of that, Cabinet Office also required DWP to pioneer implementation of its answer to the last government’s ill-fated Identity Card Scheme, called Identity Assurance (IDA). Just like that. And then, like a big red cherry atop a teetering tower of meringues and squirty cream, the Cabinet Office mandated DWP to replace its usual systems development methodology with a new wheeze, called agile, pioneer also new billion-pound contracts to accommodate new ways, recast relations with its key suppliers and invite a gabble of unknown SMEs onto choice parts of the job.
DWP proposed doing all this while undergoing a massive reorganisation, losing experienced people needed to implement the benefits reform.
It was always possible that DWP might have been able to do all this in two years. All things are possible. But anyone with the plans before them might have more reasonably concluded it was reckless.
Maude had the plans before him. He and Ian Watmore, his permanent secretary at the Cabinet Office, had helped DWP formulate the project plan. They ensured it accorded with the Cabinet Office strategy for preventing IT bodge.
Maude was personally familiar with even the technical decisions behind Universal Credit: not just what should be built but how it should be built. When DWP was in a quandary about how complex it should make Universal Credit in 2011, the Major Projects Authority – the branch of Maude’s Office concerned with preventing IT bodge – concluded that it should reach for the skies. This would be one more cherry upon the cherry atop the already teetering tower of jelly project plans that Maude’s department had signed off for Universal Credit.
The MPA gateway review of Universal Credit – the report that gave it the official green light – said DWP brass were unsure whether they should build a system to do just what Universal Credit needed it to do, or whether they should build a system that could be re-applied in any given circumstances – a sort of benefits system for all styles and seasons.
MPA advised that they should follow the latter option. This may not have seemed wise to an impartial observer. But it’s reason for doing this was political. The Cabinet Office was then laying the latter option down as part of its ICT Strategy.
“Given the Coalition Government’s desire to see re-use built into IT systems from the outset, it would be prudent to consider opportunities for this now,” it urged.
It might have been prudent from some perspectives. From the theoretical perspective of the best way to build an IT system, for example, it was prudent. This was the sort of prudence the Cabinet Office wrote into its blue-sky ICT Strategy – an inspiring work of political wonder. But it may not have been prudent to lump these new-fangled ways on a project team that already had just two years to build the impossible.
It was, said the MPA, a question of scope: “What was definitely within the Universal Credit boundary, what could be paid by the Universal Credit platform at a future date, and what was definitely out of scope.”
The MPA helped DWP design a roadmap to codify the answers to these sensible questions. This Integrated Assurance and Approval Plan (IAAP) would “ensure the correct internal and external assurance,” said the MPA Starting Gate Review of Universal Credit.
Having approved all the plans – what was in doable, what was not, what was wise, what was bonkers – MPA passed them for approval to the DWP, and to the Treasury, which would convene a meeting of the Major Projects Review Group (MPRG) Panel for a last, reassuring stamp of approval.
Representatives of the Cabinet Office, Treasury, HMRC and the Department of Communities and Local Government (representing local authorities) then became responsible for routine governance on the Universal Credit Programme Board. Watmore had a place. Iain Duncan Smith, the secretary of state for Work and Pensions, was chair.
“Cabinet Office and Treasury … ensure that what is going on is appropriate. The role the Cabinet Office can play in those situations is to ensure and quality-assure that what people are doing is the right thing,” Watmore told the Public Administration Committee last year.
Even all this careful attention could not in the circumstances have seemed to the Cabinet Office and Treasury enough to prevent Universal Credit becoming another IT bodge. But they had a magic ingredient.
Sanctioned by the Cabinet Office, DWP went into Universal Credit with the belief that agile systems development would make everything possible.
“By adopting an agile development approach, it has reduced the time it will take to deliver Universal Credit by almost half,” the Institute for Government put it in its influential report, System Error, that put wind in the sails of Cabinet Office strategy in 2011.
Mark O’Neil, director of innovation and delivery at the Government Digital Service, the branch of the Cabinet Office leading the agile initiative, was charged with helping Universal Credit directors get their heads round his favoured systems ideology.
They saw agile had a lot going for it. But it is hard to credit that these new ideas blinded them of sense. Even Superman could not rescue a cat from a tree made of Kryptonite.
Cabinet Office made DWP deputy chief information officer Kenny Robertson senior responsible owner of the cross-government agile stream of its ICT Strategy, as if he didn’t already have enough to do. It charged him with formulating a plan for agile development to be replicated across government departments, and pulling together a team of agile suppliers to carry it out. And launching a centre of excellence. And a contract vehicle.
Within months of Robertson completing his tenure as agile SRO in summer 2012, agile was kicked off Universal Credit.
The Cabinet Office appears now to be trying to save agile by blaming DWP for mishandling it.
“The Cabinet Office does not consider that the Department has at any point…appropriately adopted an agile approach to managing the Universal Credit programme,” said the NAO report on Universal Credit last week.
Yet the Cabinet Office frequently cited DWP’s agile work to flatter its own agile policy when promoting its ICT Strategy.
“Universal Credit programme is one of the first ‘Digital by Default’ services, using an Agile approach to reduce delivery risk and improve business outcomes,” boasted the Cabinet Office’s Strategic Implementation Plan in October 2011.
“Agile … the best practice that seems to be out there at the moment,” Watmore told the Public Administration Select Committee in April 2011. “I went up to see the Universal Credit programme in Warrington the other day, and that is precisely what they are doing.”
The MPA was also well aware that DWP had used a variation of agile methods, or had been applying them irreligeously. It said so in the Gateway Review with which it approved what DWP was doing.
Steve Dover, head of major programmes at DWP until last October even spelled this out at an agile industry meeting with O’Neill in 2011.
“Its a brilliant, brilliant methodology. [But] it’s not always applicable to everything. There’s some heavy legacy stuff [where] we implement changes on a 6-month basis. I probably wouldn’t apply agile to that. Because that works. I’m not going to break it. It supports the current business,” he said.
This year, when the Cabinet Office took emergency control of Universal Credit and sent in its agile A-Team, it left the legacy systems to the DWP. Would Cabinet Office’s crack team save the day? Or would they just build some websites and take all the glory? Neither department will say what they are doing in there.
When last week the NAO reported that the development had been bodged to date, it said the reason was DWP’s struggle to get a detailed design. Then without actually spelling it out, it recounted the numerous Cabinet Office initiatives that had put spanners in the works for Universal Credit.
DWP dropped the Cabinet Office’s Digital by Default strategy in May, having worked on it for two years. It’s attempt to implement the Cabinet Office-designed IDA system also failed.
“Cabinet Office decided in December 2011 that the proposed solution was too expensive and unfit for cross-government purposes,” said the NAO report.
“The department continued developing its own solution but there were delays in securing funding and finalising the tender for IDA providers,” it said.
DWP couldn’t have implemented Cabinet Office’s Digital by Default strategy if it hadn’t worked out a way to implement Cabinet Office’s IDA scheme as well. It couldn’t do online benefits administration for 8m households if their transactions weren’t adequately protected. It ploughed on with Universal Credit though. Cabinet Office continued referring to DWP in public as the pioneer of its IDA scheme.
CESG, the government computer security service that reports to the Cabinet Office, rejected these further efforts a year later, in January this year. Where was their help before now?The Cabinet Office had meanwhile rejected DWP’s IT infrastructure plan. This would have been the foundation of the finished system. Cabinet Office said you can’t have a foundation until you know what the finished system would look like.
What is left of the Universal Credit plan now is what there was of it before the Cabinet Office piled DWP to breaking point with all its other radically ambitious IT reform programmes: digital by default, IDA, agile.
How much time, effort and money had DWP wasted trying to shoehorn the Cabinet Office initiatives into Universal Credit?
When the country was suffering its worst recession for generations, and poverty had forced more than half a million people to get their food from charities; when the department of social security should have had its full attention on helping them up, it was instead tying itself in knots over a hair-brained IT project, a hair-brained reform of the benefits system and a clutch of hair-brained initiatives from the coalition ICT Strategy.
The NAO failed to ask the obvious questions in its account of this shambles last week. The important questions: how much responsibility does Cabinet Office carry for Universal Credit’s failures? How had Cabinet Office’s oversight of Universal Credit failed to forfend all its problems?
How could the Cabinet Office ever have been credited as an authority on sensible IT programme management when it was trying to pursue its own political objectives as well? Should it now be broken up, with its IT governance and procurement functions separated out where they cannot be confused by politics and do any more harm?
Cabinet Office’s broader political objectives will emerge best from this mess. It wanted DWP’s benefits system – and by extension the DWP – broken up. Conservative think-tank Policy Exchange has begun pushing for benefits administration to be devolved fully to the regions. The DWP’s days as one of the great departments of state may be numbered.
Universal Credit will be held up as another example of how big government IT projects always go wrong. Yet this one went wrong under direction from a department that forged its reputation by mercilessly barracking the IT failures of the last lot. The coalition government has lost its right to say any more of Labour’s ill-fated National Programme for IT in the NHS. It has made all the same mistakes as the last lot.
In October, when DWP is due to begin full roll out of Universal Credit, it will instead be extending a rump of what it planned for a pilot, from one site to just a few more sites. It’s molecular cohesion has been shaken as though through sustained shock therapy. It may be too shaken to suffer the stories that will inevitably now recount, relentlessly, over the years it may now take Universal Credit to catch up to where it intended to be today, its failures and the failures of its kind.