Over seven years, the head of Britain’s benefits services has pushed through a vast set of policies and reforms – but it’s been a difficult journey. His experience has clear lessons for both ministers and civil servants, he says, starting with: don’t expect the impossible. Matt Ross reports
Robert Devereux couldn’t have picked a more awkward time to take over the UK’s Department for Work and Pensions. When in 2011 he became permanent secretary of what is comfortably the biggest department in government, the Coalition government had recently taken office. It brought with it both a vast and ambitious programme of manifesto promises – including a radical plan to rebuild Britain’s benefits system from the bottom up – and an austerity agenda that would see operational budgets fall by 20% over four years.
Nearly seven years on, Sir Robert – as he is now – is retiring; and on his last day in the office, he traces many of the challenges he’s encountered back to decisions made in the Coalition’s first weeks and months.
When a party wins office after many years in Opposition, he points out, it generally arrives with a huge “shopping list” of plans and policies. These new ministers “don’t know about government or the civil service or accumulated risk – why would they? And they arrive in front of a permanent civil service which has for the past 10, 15 years been serving the other lot.”
The brakes don’t work
The result, he explains, is an “inevitable tension. And when we’re eyeballing each other and they’re saying: ‘Are you going to do this?’, it’s not a great moment for the permanent secretary to say: ‘Well, there’s a bit of learning here, and that is: don’t start with this ridiculous amount [of projects].’ Because they’ll say: ‘I knew you’d say that, because you’re still with the other lot really’.”
Devereux doesn’t have a systemic solution to this problem: following a change in administration, he believes, new ministers would often achieve more if they attempted to do less. “But that’s a brave call for them,” he acknowledges. “If you’re trying to win an election, are you more interested in having a long list of eye-catching things you’re going to do, or in the intrinsic do-ability of it?”
Even DWP’s non-executive directors, business leaders recruited under the Coalition in a bid to strengthen departmental governance, failed to persuade ministers to slow down delivery. “Pretty much each of them, within about a month of arriving, said to the then-secretary of state: ‘We’ve never tried to do this much change at once. So if you want our advice, make the programme smaller’,” Devereux recalls. “To which, generally, each secretary of state smiles and says: ‘Well that’s fine, but I’m afraid we’re on it already’.”
A crowded caseload
Making Devereux’s position still more difficult, he joined DWP in January 2011 – eight months after the election, and with a set of projects already up and running. These included major reforms to the State Pension and the management of child maintenance; automatic pensions enrolment across the workforce; the replacement of Disability Living Allowance with the Personal Independence Payment; and the creation of Work Programme services for the unemployed, delivered by external providers managed under a ‘payment by results’ scheme.
Far bigger than any of these, however, was Universal Credit. Frustrated by the endless complexities and perverse incentives of the UK’s multi-layered benefits system, new secretary of state Iain Duncan Smith (IDS) had spent years in Opposition developing his plans for wholesale reform.
By merging six benefits into one, then linking payments to claimants’ real earnings on a monthly basis, he wanted to ensure that work always paid: that people’s income would always rise if they took a job, and their benefits would be reinstated quickly if the work proved short-term. Affecting around 8m claimants and the disbursement of around £70bn (US$96bn) a year, this was probably the biggest single reform project in government – and IDS, who’d already had an ill-fated stint as Tory leader, knew his reputation depended on its success.
Eyes wide open
Even in 2011, Devereux recognised the dangers of trying to manage so many projects concurrently, and to such ambitious deadlines: “There’s a lot to be done, and there are manifestly more risks of doing it over one period than in taking forever over it”, he told me at the time. But he could not, he says now, pinpoint where those risks would arise: “If you try to do six or seven of the largest programmes in government simultaneously, then you’re constantly thinking: which of these is most tricky? And it is genuinely impossible in advance to say how the accumulation of risk is going to materialise,” he argues.
The task of managing those policy changes was made still more complex by DWP’s simultaneous programmes of organisational reforms: designed to meet the Coalition’s austerity agenda, these would see both the 124,000-strong workforce and DWP’s office space cut by a third. And here, he and his management team have clearly achieved something remarkable.
According to the annual Civil Service People Survey, during Devereux’s tenure the DWP’s employee engagement score has risen from 44% to 60% – against a rise of just three points in the departmental mean. Still more dramatically, his department’s score for ‘leadership and managing change’ has more than doubled from 22% to 48%, outstripping the civil service-wide rise by some 16 points.
Fewer, happier staff
This kind of continued rise in staff morale is, as Devereux proudly notes, “pretty much off the scale for almost any organisation” – let alone a big employer with a substantive organisational change agenda. “The people here are reporting greater pride in what they’re doing, by a long margin,” he adds. “Creating the conditions in which that is true is, I think, the biggest thing I’ve enjoyed here.”
To some extent, Devereux’s task was eased by DWP’s demographics: given an ageing workforce and a long-term cuts agenda, he’s shed 40,000 jobs by natural wastage. But if their work had simply been handed to the remaining staff, he’d never have raised morale: instead, the department has used business process reforms, customer service improvements and digital technologies to handle its caseload more efficiently.
Improved training of call centre staff, for example, means they’re better able to solve callers’ problems – helping to cut incoming calls from 100m to less than 50m a year. More recently, DWP has allocated each of its call centre staff a caseload of named clients. “Throughout our history, our staff have taken the next piece of work,” he says. “Now they’re dealing with 300 people they know.”
Last month, the department began trialling a new system enabling callers to speak directly to their case manager – “so now one of my colleagues is looking after you in the call centre, and one work coach looks after you in the job centre,” he notes. “That’s orders of magnitude better for both claimant and staff.” Then the quietly spoken permanent secretary allows himself a burst of excitement: “That’s the sort of extraordinary transformation we’ve managed in the last seven years!”
Big project, big problems
Meanwhile, many of those big policy goals have been delivered. Some have gone more smoothly than many expected: automatic pensions enrolment across the workforce; rebuilding the child maintenance system; big State Pension reforms. Others, such as the Work Programme and PIP, have hit problems (see box below).
But the biggest problems have emerged in the biggest project: the Universal Credit (UC) programme has seen £40m (US$55m) worth of IT investments written off, highly critical reports from the National Audit Office and Public Accounts Committee, and clashes between DWP and other key departments. The scheme has survived – but with its reputation badly damaged, its ministerial champion out of office, and its launch date put back by five years to 2022.
Devereux doesn’t deny that the initial timetable, set before he landed at DWP, was unrealistic: civil servants are, he points out, particularly sensitive about appearing obstructive to ministers in the early days of a new administration. And the scheme had a difficult birth – for IDS immediately clashed with chancellor George Osborne, setting the tone for a fraught relationship between Treasury and DWP ministers that would persist for several years.
“It’s pretty much a matter of record that the relationship between this department and the Treasury in respect of Universal Credit, for most of the time, was pretty frosty,” comments Devereux. “And that was for a very straightforward reason: the original design that Iain Duncan Smith brought into government… would have added billions of pounds to public spending. And the chancellor was trying to take some billions out.”
Osborne and IDS eventually agreed on a set of benefit cuts; but DWP ministers and officials were keenly aware that UC had powerful enemies, and that if the project were seen to be failing then the Treasury would try to kill it.
This, Devereux believes, helps explain why senior leaders weren’t made aware immediately as things started to go wrong in the project. “When there are people even within government who are looking for excuses to say: ‘Well, let’s not bother with that [project] then,’ then some of the subliminal messages – and the not-so-subliminal messages – coming right from the top of the system downwards are: ‘Actually, look, let’s manage this ourselves’”, he says.
Asked directly whether ministers weren’t immediately made aware of emerging problems, Devereux points out that “we were at the time running six extremely large, complicated programmes, and the extent of the involvement of the secretary of state and senior ministers with the UC team was very substantial: it was daily, essentially. And ministers were feeling: ‘Actually, this is making progress.’ In contrast, some of the other [projects], the child maintenance one, were quite difficult stuff. So we were trying to focus on the things that were on the face of it immediately difficult.”
When problems did emerge, Devereux argued, they had their roots in expectations – policed by the Cabinet Office and its Government Digital Service (GDS) – that UC would be a fully ‘digital’ system with public access via an online interface. This fitted the new approach to IT championed by Cabinet Office minister Francis Maude and the GDS team – but, Devereux says, “there’s nothing about it being online which changes the incentive for you to get a job.”
The UC model, he points out, works on the basis that “every work coach I employ can look you in the eye and say: ‘Take this job. It may only be three weeks, but for those three weeks you will be better off’. That has nothing to do with it being online. So a smarter strategy, probably, would have been to say: ‘It would be great to have this all-singing, all-dancing, modern and online. But let’s just have a look at the capacity of the organisation to build an entirely online system, frankly, and the answer is that it’s going to take a while to build that up’.”
So the DWP was short of the skills to build a truly digital platform. Worse, the system it had started to construct wasn’t designed with a public interface at its heart; and the team hit problems trying to ensure its online face would be secure. When “we tried to make it properly secure, it became more and more complicated – because it was basically a rather traditional system that we were trying to make smart,” Devereux recalls. “With the benefit of hindsight, I don’t think people had started out thinking: ‘Okay, this will genuinely be online, so the whole structure is going to [have to] be different.’”
The centre strikes
So in Devereux’s view, “the online nature was getting in the way of delivery”. And then he had some bad luck, losing two project managers in quick succession. The Cabinet Office took fright, announced that much of the work done thus far would have to be written off, and sent in a GDS team to take over the IT build. Deflecting questions from MPs and the media, IDS attempted to pin the blame on his civil servants – both via anonymous press briefings naming Devereux, and in public in more general terms.
For both Devereux and IDS, this was the most dangerous period in the project – but they survived; Global Government Forum understands from other sources that IDS ultimately concluded that, were Devereux to fall, the Treasury and Cabinet Office would probably ensure that UC went down with him.
Meanwhile, the GDS team was learning just why DWP’s own IT staff had struggled to tie a public front end to the department’s morass of multi-layered IT systems and databases. “Actually,” says Devereux archly, “it turns out that the benefits system is really complicated.”
Ultimately, the GDS team departed as abruptly as it had arrived. And DWP’s UC system was retained as the working pilot, used to test and roll out the new benefit, whilst a brand new online system was built alongside it. Five years on, that digital version is nearing completion. “So had that been the only game in town, it still wouldn’t have been developed any time in the first [2010-15] Parliament,” Devereux points out, noting that GDS first “told ministers it was all going to be done in a year, and the next time it was two years, and the next time three years…”
In the end, Devereux says, DWP has “built some rather conventional benefits platforms which have been functional; and then in parallel with that, we’ve indeed built the new Jerusalem with the digital stuff.” And the department has made this approach work – but Devereux acknowledges that it initially tried to bite off more than it could chew. “We should probably have had a rather more sober assessment of whether we were ready to try and do both things simultaneously,” he says. “With the benefit of hindsight, we shouldn’t have tried to make it all online as well.”
So Devereux believes he should have been more realistic about DWP’s digital capabilities – and more willing to challenge IDS’s over-ambitious timetable. Ministers are all too aware that they may only have a brief time in one job, he explains, “so standing back and saying: ‘Realistically, this is a two-, three-Parliament reform,’ is never the first thing a minister wants to hear. So if I had my time again, we probably should have called a spade a spade – and done one thing, and then done the other.”
His regrets only go so far, however: “Given that we ended up having to do some things differently, I’m not saying it was all perfect,” he comments. “But it is worth observing that the thing that’s now demonstrably working is all the stuff that we originally built.”
The centre cannot hold
Asked what he’s learned about the benefits and risks of central government intervening in departmental operations, Devereux points out that it’s the permanent secretary and secretary of state who ultimately carry the can for everything their department does.
“There’s an extraordinary asymmetry in accountability, so my secretary of state is accountable to Parliament for the welfare system; I am accountable to Parliament for the welfare system,” he says. Meanwhile, those charged with leading the civil service professions from the centre are “not accountable to anybody for anything; but they have views about stuff.”
When the Cabinet Office is “really facilitative and helpful” – offering support, advice and staff to assist in operational reforms such as estates and commercial – then they can, he says, provide “fantastic interventions”. For example, “we had a really good exchange with the person holding the flame for estates policy, and subsequently we got lots of help from them in the negotiations [with providers], so we all feel that the estates thing has gone really well.” The relationship here is clear: “I’m accountable, they’re helping: fantastic!”
However, he adds, “that is still – in my humble opinion – the minority of central interventions.” Though the system is still maturing, he argues, there are still too many instances “when it doesn’t work: when somebody just asserts their rights to tell me something without regards to the consequences for anybody who has got to explain it, like the secretary of state”.
Such interventions may be ill-founded, built on a poor understanding of the department’s policy challenges or legacy systems; and they add to the system overload that feeds risk into the department’s caseload. “I get a regular diet of missives saying could I do this sort of class of thing in this sort of way and that class of thing that way,” he comments, “and many of these things have merit, don’t get me wrong; but that’s another collection of things for me to manage as well as the nine different programmes which I’m passing legislation on”.
Coming out the other side
Since the dark days of 2013, says Devereux, DWP has become very good at “making sure that the quantity of work keeps in step with the amount of time we’ve got to do it in – and that’s a level of programme management that I’m tickled pink with.” The department’s ability to set sensible milestones has been assisted, he adds, because “the quality of the debate about major programmes – including with Parliament and ministers – has improved no end”.
Most of the major programmes his department was handling have now been delivered; and with 660,000 people currently receiving UC, he believes the project is past the point of no return. Whilst the Treasury’s insistence on working additional benefit cuts into the system has caused further delays – and prompted the resignation of UC’s longtime champion, IDS – he has “no doubt that the thing is going to roll out.”
The evidence is, he adds, that the roll-out will see unemployment rates fall as disincentives are taken out of the system. Researchers have tracked groups of people on UC and those on the legacy JobSeekers Allowance (JSA) system in a single area, “and it turns out that the UC people consistently outperform the JSA. We’ve done it three times now, and each time there is a material labour supply effect that wasn’t there previously.”
So the mass of huge, intertwined projects that greeted Devereux on his first day at DWP is now largely out of the door; as, after seven years, is he. “With the decision we made in October to push the accelerator for the roll-out of UC, that last big reform is on its way to completion; so all the things that were on my shopping list to do when I arrived are now in train,” he comments. “I’ve done enough to feel that I can draw a line around that.” He has no immediate plans, he adds, “but I will probably carry on working: I don’t think you can go from the extraordinary, 24/7 buzz of running a place as big and complex as this and then just sit at home!”
As he leaves, he looks back with a mixture of relief and satisfaction at a period of unrelenting pressure. “It’s not all been plain sailing by any means,” he concludes – but to get from his 2011 inbox to this point, “even with the bumps in the road, actually I think that’s not a bad result.”
“It would have literally been impossible for it all to have gone well,” he adds. “There was just too much stuff to do: it’s a miracle that it’s all happened!” On that note, he heads off to his leaving party – where he’ll be met by his old adversary and comrade in arms, Iain Duncan Smith. And as he departs he provides, boiled down to 13 words, the key lesson of his long struggle to wade through DWP’s insane major programmes portfolio: “Big, complicated things,” he says, “need to move at the pace they can move at.”
Devereux on the Work Programme and PIP
The Work Programme, which offered support and advice for people trying to find jobs, represented the biggest payment-by-results (PBR) programme undertaken by UK government – which itself has led this approach internationally. And Devereux raises concerns that the PBR approach is vulnerable to ‘creaming’ and ‘parking’: it encouraged providers to focus all their energies on those most likely to get jobs, and to ignore those who’d require intensive support.
The PBR approach creates “the incentive [for providers] to say: ‘If I’m pretty sure there’s nothing I could do that’s going to help you, I may as well spend as little as possible’,” he comments. “Some of these people weren’t getting much help.”
So PBR is, he adds, “a really top of the range mechanism when you want innovation… but the trick would be to use PBR to work out what’s clever; and when you have that bottled, then get it done” using more conventional methods.
Another major reform, the replacement of Disability Living Allowance with the Personal Independence Payment, has been widely criticised for withdrawing benefits from some claimants – with campaigners arguing that high success rates in the appeals process are evidence of widespread poor decision-making. But here Devereux gives little ground, pointing out that the proportion of cases that go to appeal is tiny.
Claimants often introduce new evidence at this point, he adds, altering the calculations made by decision-makers; and he makes no apologies for PIP’s medical assessments process. “We passionately want to give the right amount of money to people who are disabled,” he says. “But it’s a system that dispenses cash and some people will, I’m afraid, seek it… Billions of pounds of your taxes are being spent, and we’re trying to make this as objective as possible; we don’t always get it right, but we have mechanisms for redress.”