HoC 85mm(Green).tif

Public Accounts Committee

Oral evidence: Department for Work and Pensions Accounts 2019-20, HC 681

Thursday 24 September 2020

Ordered by the House of Commons to be published on 24 September 2020.

Watch the meeting

Members present: Meg Hillier (Chair); Olivia Blake; Peter Grant; Shabana Mahmood; Mr Gagan Mohindra.

Gareth Davies, Comptroller and Auditor General, National Audit Office, Joshua Reddaway, Director, NAO, Claire Rollo, Director, NAO, and Marius Gallaher, Alternate Officer of Accounts, HM Treasury, were in attendance.

Questions 1 - 56

Witnesses

I: Peter Schofield, Permanent Secretary, Department for Work and Pensions, Neil Couling, SRO for Universal Credit, DWP, Nick Joicey, Director General, Finance, DWP, and Bozena Hillyer, Director, Counter Fraud and Debt, DWP’ to ‘Director Counter Fraud, Compliance and Debt’.


Examination of witnesses

Witnesses: Peter Schofield, Neil Couling, Nick Joicey and Bozena Hillyer.

Q1                Chair: Welcome to the Public Accounts Committee on Thursday 24 September 2020. We are here today to look at the Department for Work and Pensions accounts for 2019-20. The Comptroller and Auditor General has once again, as he has done for the last 32 years, qualified his opinion of those accounts due to the fraud and error in the benefits system. The fraud and error that we are looking at does not include the state pension, but has increased to its highest ever value—£4.5 billion paid out incorrectly through either mistakes or deliberate fraud. That includes overpayments by the Department, so mistakes on both sides.

With covid-19, there has been a huge increase in people making benefit claims, which we will touch on at the beginning of this session. One thing that we will be asking about is whether controls have been relaxed as a result of that huge increase. But of course we don’t know yet the full impact of the increase in claims on fraud and error. And with furlough coming to an end or changing, there are likely to be a lot more claims. So we are going to be asking the Department today what it is doing to tackle this fraud and error.

I am very pleased to welcome our witnesses today. Peter Schofield is the permanent secretary at the Department for Work and Pensions; welcome to him. Neil Couling is the senior responsible owner for universal credit at the DWP, but he has also been leading the work on covid-19, so his evidence is particularly important in that respect. Nick Joicey is the director general of finance for the DWP, which has obviously been receiving a lot of money from the taxpayer, through the Treasury, to support all this extra work; welcome to him. And Bozena Hillyer is the director of counter-fraud and debt at the DWP, where she has been for about a year. She previously spent about 15 years, I think, at HMRC, doing counter-fraud work there.

I want to kick off with just an assessment of the basic numbers, Peter Schofield, for the claims that have come in as a result of covid. You wrote to us yesterday, or we got your letter yesterday, and it touched on some of these figures, but perhaps you could just outline the total number of claimants and say at which points you have seen increases in that.

Peter Schofield: Thank you, Chair; it’s good to be here. It is a key priority for us to tackle fraud and error, in terms of protecting the taxpayer, and we would be keen to talk through what we are doing about that. Yes, as you say, Chair, over the course of the first six months of the pandemic, we did see a significant increase in claims—something like 3 million claims coming through. Over that period, it’s about a sixfold increase, so six times as many claims as we might normally expect, and there were times, there were days, when we had over 100,000 claims in just one day. I am glad to say that, by mobilising colleagues from all across DWP and putting in place new automation and new ways of working, we were able to manage that surge in claims, so we kept payment timeliness well above 90% for much of that period. We were getting money to the people who needed that money. Many people were coming to DWP for the first time and needing support, and we were able to be there for them. I am delighted about that; I am delighted that we delivered that; I am delighted that we provided that service to the public when we were most needed.

Chair: But the challenges are certainly not over. We record again our thanks for the hard work of your civil servants and those brought in from elsewhere to help support that, but it’s not without its challenges. You obviously painted the positive picture there, so I am going to ask Olivia Blake to pick up on a particular issue.

Q2                Olivia Blake: Yes, this is a very particular issue. Obviously, during the pandemic, there will be a lot of people facing two weeks off, and many of these people will be on statutory sick pay. Could you just clarify the position on applying for universal credit in these circumstances?

Peter Schofield: In terms of applying for statutory sick pay while they are on universal credit—

Olivia Blake: No, if they are receiving statutory sick pay and they have not been a claimant before for universal credit—what should the situation be for those workers?

Peter Schofield: I might bring Neil in on that one, if that’s all right.

Neil Couling: People on statutory sick pay have always been able to claim universal credit if they are on the basic statutory sick pay—there is no top-up from the employer to that. If they have children, they probably qualify for universal credit, on top of the statutory sick pay. The statutory sick pay would be taken into account either by way of the work allowance, if they qualify for one of those, or on the taper. Two thirds of the value of the statutory sick pay would effectively be taken into account. There is no real change there for people; the system was already ready, pre covid, to take claims quite successfully. 

Q3                Olivia Blake: Is it possible to ask how long that might take for a claimant?

Neil Couling: Statutory sick pay is obviously paid by your employer with your earnings, so it will come in that monthly amount if they are monthly paid. Universal credit is monthly assessed, so any earnings in that month of the universal credit claim will be taken into account—as I say, tapered away as necessary. That is a pre-existing feature of the system. Clearly, we have seen the volume increase, but the system works in no different way post covid from how it worked pre covid.  

Q4                Olivia Blake: Just finally, we have heard talk of this £500. I understand that that will not be delivered by your Department. How will you take that into consideration in some of these claims?

Neil Couling: My understanding is that it is disregarded from any universal credit claim, so it is paid locally. For some people in the earnings distribution who are entitled to SSP and universal credit, they will actually see quite a gain from that for the two weeks that they are self-isolating, which is of course exactly what we want. We want people to follow the health guidance. If they have the symptoms of covid, we want them to self-isolate. In effect, it is on top, and once you add in the effect of the taper on the SSP and so forth—for some people, the upward gain calculation that I have seen is something like £400 for the two weeks’ self-isolation period, on top of what they would have normally got.

Q5                Shabana Mahmood: Can I ask you, Mr Schofield, about your resources? Obviously, you have had to redeploy and take on new staff as well. What is your current position, in terms of staffing coverage, to deal with the pandemic? Do you have enough resources going forward and looking ahead? I know we have an economic statement later that will make some announcements, but they will no doubt have a big impact on your workload as well. How are you managing the changing public policy situation in relation to your staffing, so that you can cope?

Peter Schofield: I remember the meeting of our executive team on 18 March, when we looked at the rising claims in universal credit that we had seen the day before and we moved swiftly to pivot the organisation so that we could meet the challenge of paying the claims. We took in something like 10,000 colleagues from other parts of the organisation to focus on universal credit claims. We then brought in colleagues from other organisations, and we also started recruiting. In that first wave, we brought something like an additional 8,000 colleagues into the organisation. Some were on a temporary basis; some of our colleagues from other Departments were lent to us, which was great. That was facilitated by the Cabinet Office. There were also some recruits in that.

As the Chancellor announced, as part of the plan for jobs on 8 July, we got additional money—£900 million, I think it was—to enable us to recruit additional colleagues into the organisation. We are in the second wave of that, which is an additional 8,000 colleagues. Four and a half thousand of those are work coaches. That is part of a process whereby we will double the number of work coaches from 13,500 to 27,000 by the end of March. That is the first part of that—the first 4,500. There are another 3,500 to work in service centres and back-office processes to help us manage claims.

The other key thing, which you might not pick out from the numbers, but which has been really important in helping us manage, has been enabling colleagues who have had to be at home to be able to work from home. We started the period with a situation in which quite a number of folk had to be at home, but were unable to work, and we managed to get out around 20,000 devices to colleagues at home. That has enabled colleagues who can work, but have to be at home, to be able to work at home. That builds us more resilience as well. When you bring all that together, we are in a reasonable place.

We are of course looking at what could happen going forward. We have looked closely at the OBR’s forecast for the economy and what that could mean for us, not just in terms of the number of additional claims for universal credit, but in terms of the rate at which they come in, because that is the key thing in being able to manage that. We are developing very robust business continuity programmes that build on what we have learnt, but also take into account the additional resilience that we brought in in the way that I have described.

Q6                Shabana Mahmood: I was just thinking, as you were saying that, about the availability of jobs that your work coaches can send claimants on to, but that is not something that is in your gift. On training, you have taken on new staff and redeployed people from other Departments. You have had to bring them in line with your own systems and your ways of working, but you have also brought new people to your organisation. What have you been able to learn and streamline in your training processes that might be good practice going forward?

Peter Schofield: That is a really good question. Bringing people in is one thing—it comes behind your question, doesn’t it? Recruiting the people is one thing, but they have to come in and be trained and inducted. One thing we have found is that universal credit is a relatively straightforward system for people to learn and get comfortable with. I can give you an example. I was in a service centre in Hastings not long ago, where we have on-site universal credit folk. We also have child maintenance folk, and we moved many of the child maintenance folk over to work on universal credit. I chatted to some of them. They had four days’ training, which did not take them into the detail, but it was enough to enable them to be effective and to feel that they could understand the things that they needed to do to process claims reliably and effectively, and they felt very comfortable about it. When I met them they had moved back to child maintenance work, but they were keen to be ready to move over again should that be needed. Obviously, it is easier for our own folk, because they know the way that DWP works.

For folk coming in through induction, there are effective programmes that we have been developing. I’m back in the Westminster office right now, but when I spoke to you in May I was in my local job centre in Redhill. I spent time sitting alongside colleagues who were new to the organisation watching their training and induction programme. They felt—quickly, I would say—confident in what needed to be done, but we are learning all the time. The key thing is to test and learn and improve our systems so that the period between someone signing up for DWP and being effective in role is as short as possible.

Q7                Shabana Mahmood: Finally on this, whether the staff are old or new, how confident are you that you are picking up errors? I have in mind particularly the sorts of errors that lead to underpayments, which I often have to deal with with my constituents. How quickly, even in this pandemic phase, are you able to pick up system-wide data that tells you where there might be a problem so that you can prevent those errors from being made, or can correct them quickly?

Peter Schofield: The Committee may well focus on overpayments in this session on fraud and error—I understand that—but equally important is underpayments and getting that right. I am sure you will ask me about universal credit overpayments, but universal credit underpayments are now falling and are down at 1.1%, which is a testament to the system from that point of view. If a colleague is unsure, they have people around them to help them, and that is well set up. Also, for colleagues working from home, they have a connection through Skype to enable them to quickly seek support from more experienced people. We have our enhanced checking service, which I am sure we will talk more about in the context of fraud and error, but we have ways for folk to refer on cases that they are unsure about or perhaps suspicious about.

We also continue to do a lot on underpayments to make the system as straightforward as possible for people to share changes of circumstances and anything that they need to share with us—making that as clear as possible. We have done well on areas such as universal credit, but there are some areas that we need to do more on, such as some of the disability benefits, but you may want to come on to that as well in the context of underpayments.

Q8                Shabana Mahmood: Mr Couling, did you put your hand up?

Neil Couling: I was going to add one thing, which is that we have additional sets of eyes and ears in our stakeholders. Last time I was giving evidence, there was some criticism of our relationship with our stakeholders. Since then, we have taken those on and they are really helpful in pointing out where they are seeing trends. Child Poverty Action Group have an early warning system that they share with us. Citizens Advice are also very good at that. We are open to that and if we are missing something through our monitoring, we are not too proud to say, “Well, what is going on there?” and we will take a look at the specific problem. New schemes are being developed and deployed—we may hear some stuff today about that—and we want to make sure that we get briefing quickly to our staff so that when customers come in, they are ready to deal with that.

Q9                Peter Grant: A question for Mr Schofield again. Depending on what the Chancellor announces later today, we may well see another massive increase in benefit claimants over the next few months. A lot of them will be people who have never used the benefits system before and a lot of them will be people whose employment pattern is quite complicated. What are you doing now in anticipation, first, of that increased workload and, secondly, of the consequential risk of increased fraudulent or erroneous claims?

Peter Schofield: Obviously we need to be ready for whatever happens. Our business continuity plans were absolutely tested back in March, as I indicated earlier, and they stood up. There are things that we have been doing, as I said in my answer to Ms Mahmood, to build resilience. We are confident about that in terms of the numbers, but we are ready and looking at all potential scenarios, of course, because you would expect us to do that.

In terms of different people coming to the DWP, which I think lay behind your question as well, Mr Grant, and them perhaps not being familiar with the benefits system, one of the things that we have noticed is that to some extent we are seeing a different range of people approaching us—more self-employed people was one of the early patterns that we saw back in March, for example. We are finding that many folk who are new to universal credit are engaging and finding the universal credit online digital system quite an effective way of engaging with us and finding it quite intuitive. Because it is based on a lot of user research over the years it has been in development, it is more akin to the sorts of processes and systems that sometimes people are used to in other walks of life.

In terms of the detail of that, Neil may want to say more about some of the specific work that we are doing around user research and developing our systems so that they are as accessible as possible to different groups of people. Neil, do you want to add to that?

Neil Couling: I was going to add two things. First, it is a bit different from March. I was quoted in The Sunday Times a couple of weekends ago saying that the claims spikes were almost generated by the announcements of the schemes coming out of HMRC, because quite necessarily for my colleagues there, to put those in place, there was a lapse between the announcement and the scheme. We saw that particularly from the self-employed where, three days after the Chancellor’s announcement of the self-employment income support scheme, the claims share for the self-employed was about 13% of the caseload, so we saw a big surge there.

We are in a different situation today. I cannot say exactly what the Chancellor will be announcing, but we have the notice and schemes are in place already, so we might not see the kind of claim spiking that we saw before. That was unprecedented, even in the three recessions I have been through professionally, when claims were going up fast, but nothing like the kind of spike we saw back in March. I do not think we will see that again. If we do, we will respond but, as Peter has laid out, we are in a much stronger position than we were back in March.

In terms of listening to our claimant feedback, there is a whole development feedback loop going on.

Q10            Peter Grant: One of the things I noticed from a lot of my constituency casework was, especially, people who had been self-employed running their own business who, when they eventually had to stop trading, were so unfamiliar with the benefits system that they did not realise that they could claim benefits until a significant time afterwards. What flexibility have you been able to exercise so that people could claim benefit from the point when they should have claimed, rather than from the date that they realised they were entitled to claim? Bear in mind, as I say, that these people are not familiar with how the system operates.

Peter Schofield: There is some discretion, of up to about a month, to look at backdating. That has not changed through this process.

Alongside that, our folk are absolutely there to help people make claims. I have been sitting alongside colleagues in my local jobcentre as they made outbound telephony calls to people. We were very much in a situation of saying, “Don’t call us; we’ll call you”, and we went through them methodically. I didn’t listen in to any of the calls myself, but I talked to colleagues in the tea room after they had had a morning of calls maybe, and they said that they sensed the feeling of support from claimants, the sense of relief that someone from DWP was phoning them up and helping them, guiding them through the process. Maybe that is rather new for us, in terms of engagement, but colleagues felt that they were doing a great job helping people through.

We have had similar feedback from the other side, from people who have felt supported on the way through. But in terms of the discretion to backdate, we have not changed that—it is up to a month, as it has always been.

Q11            Peter Grant: Thank you. I have one further question, specifically on covid consequences. Last week, my colleague Chris Stephens discovered through a parliamentary written answer that in May 1.6 million benefit claimants had had amounts deducted from their claim, most of them because of previous advances of universal credit. The figure for my constituency was 3,800, and we were by no means the worst affected constituency in the UK. Mr Schofield, how comfortable are you that that was an appropriate action to take at that time, given the circumstances that everyone was dealing with?

Peter Schofield: The policy on advances deduction is as it has always been. Alongside that, though, we had a period of pausing debt collection more broadly across DWP.

Alongside deductions, the other element to this, which you will be fully aware of, Mr Grant, was the increase in the standard allowance within universal credit of about £20 a week, or just over £1,000 a year. That has been a key thing that has taken people through. To a large extent, that would have offset deductions that folk would have had, and I know it has been widely welcomed.

Q12            Peter Grant: As far as deductions are concerned, am I correct in taking from your answer than any decision not to make those decisions would have needed direction from Ministers? It was not a decision that was in the authority of your Department without ministerial instruction.

Peter Schofield: Yes. Policy on deductions is policy for Ministers. In terms of advances deductions, which is your specific point, that policy didn’t change. Neil, did you want to add to that?

Chair: Very briefly, Mr Couling.

Neil Couling: I want to be fair to Ministers here. If, hypothetically, we had decided to interrupt advances, that would have required me to de-automate an automated part of the system, and that would have put extra stress on the system. The outcome might have been that we weren’t able to pay the 3 million claims that we had in that period. That is one of the things you are weighing up: you are not considering things in isolation, but looking at a whole-system approach. It was the automation and the ability to work at great volume that got us through that unprecedented claims spike.

I was, quite aggressively I suspect, arguing in various quarters and in public—I said this to the Select Committee—that if you press for lots of policy changes, in theory you could do them, but the only way you can do them is make them manual. The thing we were short of at that point in time was people to clear claims. That would have been an extraordinarily difficult thing to pull off. You are right that ultimately it was a policy choice. I would have been saying to Ministers that, had they wanted to go there, that would have created huge disruption in the administration of the system. 

Chair: We recognise that there is a big challenge. Policy changes on the hoof can have huge logistical challenges, particularly with the logistical challenge of covid in general. I will ask Gagan Mohindra to come in, and then we will move on to the main session shortly after that.

Q13            Mr Mohindra: Following on from what Mr Couling was saying, may I put on the record my thanks to the Department? The system reacted very well, given the 3 million-odd claims that we have had since March, so well done.

My question is about the Budget. We are expecting the Chancellor to report to the House later this afternoon. Mr Joicey, how have you adapted to that? Obviously, we are potentially going to have a significant further use of UC, and access to your Department later next month when furlough ends. Are you bidding in to the Treasury, and what is the response?

Nick Joicey: We are working very closely with the Treasury, and have been working very closely with it since the start of the pandemic. We have been able to agree additional funding with it. As Mr Schofield said, right at the start of the pandemic we were able to agree some additional funding to bring in additional people. We then worked with the Treasury in the run-up to the Chancellor’s summer economic statement, and we agreed an additional £900 million to double the number of work coaches. That was the funding to enable us to increase the work coaches by 13,500 and recruit additional people.

We also announced the launch of the kick-start scheme and agreed £2 billion of funding for it. We have also been able to work closely with the Treasury, in terms of the other impacts of covid-19, not least how we ensure that our estate—our buildings—is safe for customers and people working. We have been able to work with the Treasury to secure the funding for that.

Overall, we have been able to work closely. That clearly reflects the emphasis on the importance of public health and the economy, in terms of ensuring that we are able to pay people promptly and on time, and ensuring that we are able to provide that employment support for people during this period.

Q14            Mr Mohindra: Thank you for that. Let me move on to Mr Schofield. I want to focus on the fraud and error risks. As you will be well aware, the NAO has continually highlighted that as a problem within UC. As per the latest report, there is a suggestion that there may not have been a priority to target fraud and error. What is your view, and how do you give this Committee reassurance that we are on top of this?

Peter Schofield: Thank you, Mr Mohindra. Tackling fraud and error is a key priority for us, as I said in my letter. I recently wrote to the Chair to say that very clearly and put that on the record. The current level of fraud and error in DWP benefits is higher than I would like, and I am sure we will get into what we are doing about that. I should step back and remind the Committee that of course, we always knew that as universal credit took over more of the working-age population in terms of benefits—that is, the population of tax credits are generally migrating to universal credit—there would be upward pressure on the measure of DWP benefits. As the overall proportion of DWP benefit claimants becomes more weighted towards that working-age population, the nature of the eligibility criteria for the benefits and the nature of people’s lives is such that there is more scope for overpayments and to get things wrong.

We always knew that there would be upward pressure. I am glad to say that, as the NAO has recognised in its Report on getting to first payment, the overall Exchequer loss across this population is reducing. We have the numbers for 2018-19, so we are seeing a small reduction; we are seeing some of those savings starting to come through. However, we have not done as much as we would want, and on universal credit, we know we need to do more. The reassurance for the Committee—I am sure we will get into this—is that we know what is causing that set of overpayments. When you look at the universal credit fraud and error numbers, the overpayment numbers for last year, 2019-20, came in at around 9.4%. We could take you through what elements make up that 9.4% in a bit of detail and what we are doing about them, so we know what we are doing, but the numbers are too high.

Q15            Mr Mohindra: Just to interrupt you there, what is the 9.4% in pounds and pence?

Peter Schofield: That is £1.7 billion.

Q16            Mr Mohindra: So, a significant amount of money. I appreciate the challenge, but the PAC rightly focuses on value for money. Obviously, the Department has had a significant workload in the past six months, but a criticism that the NAO has made—in my view, correctly—is that this has been an ongoing issue. The direction of travel is good, but when you are talking about billions, my residents would rightly be querying why we cannot get better. I look forward to seeing future iterations of the NAO Report that suggest we are fully on top of this.

If I may move on to the savings side of things, Mr Couling, there were savings targets when UC was starting to be rolled out, specifically to do with fraud and error. Will those savings ever be delivered?

Neil Couling: One of the things that the NAO says in its Report is that, for what I think is the first time, the Department has actually got the scope now to properly tackle fraud and error loss. I have been around for a long time, as the Chair will tell you—there is a nasty joke on the executive team that the accounts first got qualified when I joined the Department, which I think is correlation and not causation—so I have seen this come and go, and I have seen attempts to deal with this. For the first time, we actually have the tools to do the job. You have seen this in the pandemic: we have had to focus on attacks on the system and developing responses to them, which will see us not just through the pandemic, but through the future life of universal credit. We have the tools; we just need the space to apply them.

Occasionally, the Secretary of State calls me the abominable no-man of the Department, because I am always saying, “If you do that, you cannot do this.” What has been happening a bit with the fraud and error agenda—this is what the NAO is reporting—is not a lack of focus, I think, but an opportunity cost; there are choices to make. One of the things the pandemic has done is to force fraud and error response further up the agenda, which I think is a good thing for the integrity of the system and, ultimately, public confidence in it. We have the capability now to do much of this, and Bozena’s addition to the Department is really helping us there, but we are not where we want to be yet, as Peter was saying. I am not happy with the figures, and they do not yet achieve the business case savings. We are well off those at the moment, but we can get there because of what we have built. We just need—I hope I might occasionally see the NAO nodding at this point—to apply that now.

Chair: I think that moves us neatly into the main part of the session, which is to really grapple with the day-to-day business of the Department. It is difficult at this point to separate it from your covid claims, and the businesses of the Department is, in effect, shifting because of that volume, and the work coaches and so on—the role that you have in helping people getting back into work—but we want to get into the detail because we know that underlying the current situation is this fraud and error problem. As you highlighted, Mr Couling, it has been 32 years in the making.

I am going to ask Peter Grant to kick off. For anyone who is watching, we may be referring to the DWP’s annual report and accounts and the figures in there. Over to you, Mr Grant.

Q17            Peter Grant: May I come back, first of all, to the comment that was made a few minutes ago about the business case? The business case for introducing universal credit was largely based on a promise of reductions in the cost of fraud and error. Why has that not happened yet?

Peter Schofield: Shall I come in on this, to begin with? This is a big driver of the overall numbers that we are seeing for fraud and error in the Department. Overall, when you look at welfare as a whole, there is a reduction, because we are delivering savings compared with the position on tax credits when you look at the overall Exchequer loss, but that is no excuse. It comes back to understanding the causes of fraud and error in universal credit and where they are coming from, and which bits of the business case we are delivering on and which are the bits we have more work to do on.

If you take that 9.4%, traditionally the biggest area of fraud and error in income-related benefits has been having the wrong income going into the system in the first place. The big saving that we knew would come from universal credit was using the real-time information on earnings from HMRC and bringing that across into the automatic way in which we calculated benefits. The income from employment through the RTI feed—that contribution to the 9.4%—is now just 0.3%. That is a big move from universal credit, and that element, which is a big element of the eligibility criteria for income-related benefit like universal credit, is now down to a very low level.

That is what we were aiming for, and that is what we have delivered, but there are other types of income where we know we need to go further. As we have been working this through, it has become clear that there is more fraud and error in self-reported earnings than had been anticipated. We are doing a lot of work—we can talk more about this if you like—with HMRC to have more information around self-reported earnings and earnings from self-employment that enable us to get the same improvements as we have already seen in earnings from employment in other types of earnings. That is one big area of focus.

A second area of focus is around capital. In tax credits there are no capital limits, and 2.8% of the 9.4% relates to errors because of undeclared capital. If I was being kind to myself, which I might be but I doubt you would be, I would say that that is an additional challenge, because it is something that was not there in tax credits and is there in universal credit. But that is no excuse, and we are looking very carefully, particularly with colleagues in other parts of Government, to see what data we can use to get better access to capital information—for example, using information that HMRC has on income from interest. We might be able to talk a bit about some of the work that we are doing on that.

A third area we have to make further progress on is around what we call living together fraud. It is the fact that people’s relationship situation changes and maybe they do not declare it. Obviously, at the heart of this, as with self-reported income and capital, is helping people to know what they need to do, so that they tell us about changes of situation up front. A lot of work is going into making that more straightforward on all these different elements. We are also looking at other types of data matching that we can do to understand changing relationships and how that plays into eligibility.

It goes back to what Neil and I were saying earlier. We are not where we want to be, but we know what we need to do. That is the nature of the conversation we are having with the National Audit Office, which is rightly challenging us on this—“In that case, come up with a strategy in each of those areas, design it into the system, measure it and evaluate it, and then go back and redesign.” That circle of driving out fraud and error in our systems is the thing that I believe will lead us, over time, to a point where we can satisfy you, as Parliament, and through the NAO and others, that we have got things under control.

Q18            Peter Grant:  Very briefly, with hindsight, was the business case too optimistic in its predictions of reductions in fraud and error?

Peter Schofield: No, because we are going to get that. That goes back to what I was saying. I am confident about this, because it goes back to what Neil said about having the tools for the job. We now have to develop them in the way I described. Neil is always balancing the need to manage other challenges with universal credit—for example, managing 3 million claims over this period and the need to make the system as resilient as possible for higher volumes—with priorities around fraud and error. We know what we need to do. We need to balance that and get it delivered. We are confident that we will get those savings, not only from fraud and error but from the economic benefits of encouraging people to look for work, as in the business case.

Q19            Peter Grant: When do you expect to get to the level of savings anticipated in the business case? By the time you get there, what will be the additional costs to the public purse of not getting levels of fraud and error down to the business case figures by the time the business case said you would?

Peter Schofield: We want to get to the level of the business case, but I want to go beyond that. I want to get to a point where I am the first permanent secretary in 35 years to not have my accounts qualified. The National Audit Office are quite rightly challenging us and holding us to account on your behalf, as you would expect, but they are also working with us to try to make sure that, together, we have that journey mapped out and that we know what we need to do to get to a cost-effective set of controls. Ultimately, the same tests that they would apply are the same that you would apply and that I would apply.

We are an organisation whose job, quite rightly, is to provide support to people who need it. That is fundamentally what we are trying to do. There is always a balance here, but we need to get to a point where we have that balance set in a cost-effective way, so that you have confidence in the regularity of our payments. There will always be some fraud and error, I regret to say, but having that at a cost-effective level is consistent with our wider objectives of providing the service that our citizens expect.

Q20            Peter Grant: With respect, Mr Schofield, the question I asked was when you, as the person in charge of the Department, expect to deliver the levels of savings promised, which was the basis of introducing this system in the first place. Is your answer to that that you do not know the answer?

Peter Schofield: Coming into this year, my plan was that we would have a target set for the organisation on fraud and error, and then we would look at setting a path to go beyond that. The reality is that we do not yet know how the context we are in will play out. Many more claims naturally means more fraud and error in the system. Giving you a detailed forecast about what is going to happen is impossible to do well. I do not know what the baseline is, to be honest, because we do not know what is going to happen; we know what has happened over the past six months, but we do not know what will happen in the next six months. I hope the Committee will understand the reluctance to give a specific answer to that. It is not normally my approach to try to hedge around. In this particular case, I do not know, but I can assure the Committee that we are working with the National Audit Office and others to make sure that we are developing cost-effective controls and, just as we have done for earnings from employment, which we have got down to 0.3%—to where the business case wants us to be—we will get there on everything else. Some of that requires other powers, potentially—information about bank accounts, for example. We may need more legislation on that front to enable us to do that. Those are things on which we have been working closely with other Departments and with the Government more broadly.

Q21            Peter Grant: The Committee has previously recommended that the Department set targets as the basis for action to reduce the levels of fraud and error, whether those targets be in monetary terms or expressed as a percentage of the number of claims, which I appreciate might be more appropriate just now, as the number of claims is increasing significantly. Can you tell us what progress you have made in implementing those recommendations to start setting targets for your organisation?

Peter Schofield: We were all set to do that, and we were going to propose a target of around 2.3% for the year that we are in, but we have not because the effect of covid is such that we need to see how that plays out in practice. As soon as we have a clear baseline, we will go back to setting a target, working with the National Audit Office on how we do that.

I absolutely accept in principle that we should have a target; I think that is right for an organisation like ours, given the level of public scrutiny and interest in the question of fraud and error in the DWP and benefits. I accept that in principle, but in these quite extraordinary times, knowing where to set the level is quite difficult. We will come back to that when we have a clear baseline.

Chair: Mr Couling, I am aware of time, so if you could add the bits you want to, that would be great.

Neil Couling: I will be very quick, Chair. I just wanted to add that, of course, covid brought us two challenges as well. Our traditional measurements system did not work for covid, because when we sample, if people do not contact us, we can stop their benefit entitlement to allow us to sample, but we did not want to do that in the middle of a health emergency, so we had to abandon that. In fact, I snaffled the investigators to come and work as part of the 10,000 on the big surge of claims. We have moved those people back now, and we have restarted our measurement of fraud, but not only has covid knocked all the claims numbers all over the place, it has made measurement a bit more of a tricky issue for us. We will, however, produce the best estimate we can, subject to further developments on universal credit given that, as Mr Grant was saying, I am keen to achieve the business case benefits here.

Q22            Peter Grant: Is it your intention to set targets broken down by individual benefit type? Quite clearly, some benefits are more vulnerable to some kinds of error than others. Is it your intention to set targets for different kinds of benefit, or will it just be a single umbrella target for the whole organisation?

Peter Schofield: We have been looking at that. Our plan was to have an overall target for the year that we are in, but that is something that we will keep under review. This is something that we talk to the National Audit Office about. What is the point of a target? A target is to incentivise us and to give a degree of public confidence and transparency in what we are trying to do. We will work with them on that to get it right at the point when we have clarity about the baseline to know how to set it up.

Q23            Peter Grant: Neil, you are going to be set some hopefully quite ambitious targets. Do you have the resources just now, or do you expect to have them once covid has settled down, to allow you to meet those targets?

Neil Couling: There are two sorts of resources. There are my resources: the people who build the system, who are in place and doing that. Then, there are Bozena’s resources, who both inform that through their intelligence work—what should we build?—and then conduct compliance activities, fraud investigations and the like, to challenge that.

Q24            Peter Grant: Thank you. I will come to Bozena, if I may. Have you requested additional resources to what you have already got? Are the resources you have just now sufficient to meet the targets you are going to be set?

Bozena Hillyer: The resources that we have at the moment we have used really going back to the NAO principle about value for money, and putting resource where the biggest risk is; so in terms of the response to covid, a number of my people went to work on the frontline, but we kept in place enough capability to make sure that we were absolutely monitoring and understanding what was happening, reacting to anything we saw in the system, and putting in place new and different types of teams to be able to disrupt, for example, in the serious organised crime space, and to do enhanced checking.

Obviously we are, at the moment, going through the spending review and the spending round. Our strategy, in terms of reducing fraud and error, is to prevent it where we can, and then, where we can’t, to detect it as quickly as possible—and then to correct and collect as efficiently and effectively as we can. In terms of prevention, what that means is really not allowing anything in the first place.

When we talk about resources, much as Neil was saying, it is a balance between people and technology. What we are building, and what we have got money for currently and are starting to build this year, is a system that is based on transaction risking. Our vision is to be in a place where we can in real time, or near real time, assess every claim as it is coming through and take a view of how much we trust the information that is in the claim, and have different customer journeys.

This is where working with Neil’s people in terms of the programme and also operationally is so important, so that we have one customer journey for people where we trust the information and it is fine, and it’s as quick as possible and as easy as possible, because fundamentally that is what we are here to do—to pay those people; and a different customer journey with more intervention where we don’t trust that information.

In terms of overall resources we will be asking for resources both for people and for investment, as will, I am sure, lots of other Departments, in order to make sure that we can actually make our strategy come to life.

One of the things I would say, and a reflection on what we found going through covid, is that the pace at which we have been able to do things has accelerated our learning on what is possible. We really do see that putting more investment into data, into data analytics and into that prevention space, is going to get us where we need to go, coming back to your previous question to Peter, about when we will get to a place where we can reduce the levels of fraud and error. That is the direction of travel. So yes, we have got the resources we need at the moment; yes, we will be making bids for the resources that I think we need to deliver our strategy.

Peter Grant: Thank you. I appreciate, Chair that a number of Members are quite keen to get in, so I will leave it at that. I may want to come in at the end, but I suspect we will be quite pushed for time this morning.

Q25            Chair: I just wanted to come to Mr Joicey on all these targets. Mr Joicey, you are in negotiations, locking horns with the Treasury for the spending review. There has been a lot of positive talk from the permanent secretary about having targets for fraud and error broken down by benefit—all the things that we have been pushing on for a while. When you talk to the Treasury, how much are they pushing you on this, and are they tying your hands? You would expect them to tie your hands, with all the money they have given you already to improve on fraud and error; and you are the finance director. It is your accounts that are qualified. How does that make you feel? A lot of questions there. Sorry, I am trying to be quick.

Nick Joicey: First, this is enormously important, and that’s why as finance director general it is very important to me that we are able to have this hearing today with you focused on our annual report and accounts with the figures and information in front of us that enable us to drive progress. I should say on your comment about locking horns with the Treasury that we have been having very constructive discussions with the Treasury and we have worked with the Treasury throughout the pandemic. I spoke a bit earlier about the resources that the Treasury has made available to us. As you would expect, those resources are there to ensure that we can deliver what we need to deliver as a Department, in terms of making payments to people and being able to support people into employment, and alongside that, as you rightly highlight, ensuring that we are paying the right amounts to the right people and addressing fraud and error.

In the context of the work and the further discussions that we will have with the Treasury on the spending round, I think it picks up on Bozena’s point about the innovation that we have been able to demonstrate during the pandemic, in what are clearly very difficult circumstances. For example, we have been able to automate some processes for some of our other benefits, to deal with the additional pressures. And the Treasury are very interested in working with us on how we can build on that innovation and to fund that investment in order, as you rightly say, to bring down fraud and error.

Q26            Chair: That sounds very pally, but are they not going to be a bit tougher on you on the targets, as money is pouring out—rightly—to support people who are out of work? Surely they are being a bit tougher than that. I mean, have you actually had discussions about timelines for reaching targets on fraud and error, because that is millions of pounds to save that can be spent on getting people into work? I am sure the Treasury is watching it in the same way we are.

Nick Joicey: We work very closely with the Treasury. And regarding your comment about whether it sounds pally, what we did set up and what we have been doing is having joint teams with the Treasury. So in the run-up to the summer economic statement, we had a joint DWP and Treasury team working very closely together, and as you would expect, there is a lot of interrogation of our plans and assurance, in particular, on delivery.

With regard to fraud and error, the Treasury have wanted to see very clearly what the risk judgments are that we are making in terms of being able to pay people on time, alongside the risks of fraud and error that other colleagues have highlighted. So as you would expect, the Treasury absolutely have been looking at that very closely with us.

Looking out over the spending round, I expect that we will be able to have those discussions with the Treasury and make progress with the Treasury, in terms of getting investment in fraud and error. However, as you rightly highlight, that will be linked to what we are able to deliver in terms of benefits for the taxpayer, regarding bringing that overall expenditure down.

As for the actions that we have taken so far, our estimates are that the actions we took prior to covid are saving us £1.5—

Q27            Chair: I will go to Mr Mohindra in a moment, but you have got a challenge. You have got the baseline of your budget, with the fraud and error that we have gone over many times in this Committee, and you have got the whole issue of covid and all those extra claims, trying to get a baseline that you, as finance director, can keep an eye on, measure and prove to the Treasury, the taxpayer and this Committee that you are watching and bearing down on fraud and error, which is costing the taxpayer dear. How will you make sure that you can keep that measurement, so that we can all see very clearly and transparently the line through from pre covid, through this difficult time, to a period when things may settle down in future, so that we can have a good baseline to measure against?

Nick Joicey: It is enormously important that we have got that baseline. Mr Couling referred earlier to work—I think the phrase that he used was that he had ensured at the very outset of the pandemic that we were mobilising people to make payments. However, as you would expect, for me, as the finance director general, it is absolutely critical that we have in place the processes and the people to be able to measure fraud and error, and to ensure that we have that baseline.

Clearly, at the very start of the pandemic, we were not able to undertake the measurement work we normally do, because that involves visits and detailed sampling, which we were not able to do right at the start of the pandemic.

In July, we started that work again on universal credit, so we are undertaking the survey on universal credit, and also on other benefits, including pension credit, ESA and so on. From my perspective, it is vital that we do our absolute best to have a measure and an estimate of fraud and error that we can include in our annual report and accounts, to the usual timeline, and to be held to account by yourselves and others.

At the same time, we need to recognise that, as I say, the pandemic and the impact in particular in those very early months clearly has an impact on what we can deliver this year, but we have to make our best efforts to do that, and that work is happening now. We have redeployed people back to that work.

Q28            Chair: A final one from me, which may be to you or to Mr Couling. It is about whether you are separating out some of the measurement of the fraud and error in the static claim—the normal workload—from the newer claims that are coming in from people in different types of jobs and people who have not been interacting with the system before. Are you looking at that, in order to try to keep that baseline measure so that you can track it?

Nick Joicey: I will pass over to Mr Couling in a moment. Our approach, guided by advice from NAO colleagues and others, has been to, as best as we can, replicate the sampling that we would be doing in a normal year, ensuring that we can get as close as possible to sampling the same proportion of benefits and the same proportion of claims within benefit areas. As I say, there are constraints on what we can do, but that is our approach.

Peter Schofield: I will build on that. One of the things that I think I have shared with the Committee before is our concept of something called a heat map. It is not just a high-level number; it is looking at the causes of loss, benefit by benefit. You end up a with a matrix—a grid—which is something that we can then focus on, with an individual strategy for each case. There is also a whole load of profiling that is being done by Bozena’s teams, in terms of understanding the different nature of risk in different cohorts and the different journeys that they have come through.

All that then plays into the management information that we have, as opposed to the things that we would publish, because this is about management information in real time that we can use to develop our approaches. I would not want to go into more detail on that here, but it is the thing that gives me reassurance as accounting officer that we have strategies that are driving down fraud and error. It is not just at a high level, because it is quite difficult to develop a strategy across the board. It is a more sophisticated approach that gets into the different types of loss in different cohorts, enabling us to focus on each of those.

Chair: Thank you. I will now ask Gagan Mohindra MP to pick up.

Q29            Mr Mohindra: My question is to Mr Couling. Reference was made to RealTime Information, which is data provided by HMRC. Unfortunately, not as many claimants are picked up by that system, especially those who are self-employed, who work in the gig economy or who work for cash in hand. How are you tackling the fraud and error among those without an RTI return?  

Neil Couling: I might hand over to Bozena for some of this, because my focus is on what I can automate and how I can build systems that provide information.

Mr Mohindra: I will ask Ms Bozena Hillyer to answer that question, then, please.

Bozena Hillyer: You are right: RTI is pay-as-you-earn, and that is where people have earnings from employment. We also have other datasets from HMRC, because everybody has to make returns, in terms of what their income is—whether it is from self-employment or other self-reported income. The main issues we have around this are more about time lags. We have data. That means we need to supplement them with data from different sources as well. I do not want to go into too much detail about what those might be, but they would be the obvious things: different agencies and financial institutions or financial companies that would have information on people.

When we are doing an individual investigation, we actually have the powers to go and get the information directly from a bank or whoever else it is that we might want to approach to get the information. When we are looking at that bulk place—this is why we have a challenge here, which was alluded to earlier around the UC business case—we know we are doing well on the RTI part, but we know we have more to do on the self-employment part. It is because of the time lags in the data, and that means we have to look for other data in order to be able to match. The golden thing here would be to have a dataset that could match directly. We work across Government, and outside Government in terms of things such as the joint fraud taskforce, in order to source and use that data. We also use different pieces of legislation, such as the Digital Economy Act, to be able to get us those datasets that allow us to match against things that do not form part of that direct RTI feed.   

Neil Couling: If I may add one thing from a historical perspective—Nick Timmins writes about this a lot in his brilliant assessments of the welfare state and its development. The self-employed have always represented a challenge here, so I would not want the Committee thinking we have got this nailed. This is going to be hard work, and the growth in self-employment is one of the reasons why the business case numbers are a bit askew at the moment, as everybody is pointing out. I have got time on the business case: I am meant to realise all the benefits on that by 2027-28. We are going to apply our best endeavours to this task, but it has been a challenge on tax credits and it was a challenge in the pre-tax credits world as well.

Chair: I think that we all know that covid has thrown up the challenges of different types of work, which hits your Department particularly hard.

Q30            Mr Mohindra: Reference was made earlier to the difficulty about claimants’ undeclared capital. Ms Hillyer, what are you going to do about the lack of data, and, as a follow-up question—perhaps Mr Schofield will want to come in on this—do you need more regulatory powers to address this challenge?

Bozena Hillyer: In terms of capital, as Peter said earlier, it’s the same as the self-employed problem; the question is: what datasets do you have that you can match to what a particular claimant might be telling you? Again, we do have data flows there, but there are time lag issues. As I alluded to earlier, we have powers that allow us, when we are doing an individual investigation, to ask for all the information that we need. Where we may well be moving to is looking at whether there are more powers that we need—but not just us. This is really a cross-Government thing, and we are currently working with the Cabinet Office and with other Departments like HMRC to look at these questions. What are all the powers that we have? Are there gaps? Are there things that we need?

We are always very cognisant as well of the fact that we might want to ask for information, but there are sensitivities in terms of privacy on the other side. But we have done a sort of proof of concept around using, for example, banking data, which gave us really good results. So yes, one of the things that we would be looking at, and actually are constantly looking at, is this. Is the legislation right? Do we have the right amount of powers for the way things are moving and the new types of crime and new types of attacks that we see on our system? That is definitely one of the things that we are looking at in terms of our response to the question: how do we crack this challenge we have around capital?

Peter Schofield: Let me just add to that, because I think it is an example of a broader theme here, which is where we take our understanding, from a heat map or wherever else, of where the causes of loss are. We then develop an approach, look at datasets that we might have access to, and then do a test and learn. Then we work collectively with other Departments; Bozena has talked about the cross-Government approach here of working closely, brought together by the Cabinet Office. And then we can make a case to roll something out at scale, in whichever form that is. That could be legislation. In some cases, it’s automation of something. But it is a question of test, learn, apply and measure/evaluate. It’s this constant circle. This is the thing that will get us to a point where we can drive this out.

That’s one example. It’s a great example; it’s a really important example, because it’s 2.8% of the 9.4%, so I really want to get progress on that, moving forward.

Q31            Mr Mohindra: I am sure I speak on behalf of the Chairman when I say that we look forward to hearing whether you feel that there are some bottlenecks that we can remove as a Committee—I have to be careful about going any further than that.

Peter Schofield: Your offer of help in advance is much appreciated.

Bozena Hillyer: I just want to say, to reassure the Committee, that doesn’t mean we are not doing anything. We are constantly looking at the data that we have; we produce scans to identify where we think the highest risk is; and we have teams that have already started and are ready to go, to actually put that right. So in the absence of this, we are not doing nothing. We are, absolutely, still doing everything that we can with the information and data that we have.

Q32            Mr Mohindra: That is good to hear. I think that, from the Committee’s perspective and from the NAO’s, it was an obvious dataset that historically you may not have had access to, and we are just mindful that, where possible, we want to be a helpful friend—so use the opportunity in the future.

If I may, I will move on to the overpayments. Mr Schofield, how are you looking to reduce the £1 billion overpaid to recipients of disability benefits because their health has improved?

Peter Schofield: That is an area where there has been some improvement. I have to say that, on disability benefit, although there is something here around overpayments, there is also a really important focus on underpayments. Indeed, on underpayments, disability benefits are the biggest cause of that. There is a real focus here on a number of things. Bozena might want to add to this. A lot of the reason for getting it wrong is because people’s conditions change. In some cases, happily their function has improved. That is what leads to overpayments, so we need to be able to pick that up and encourage people to share when that is the situation. Of course, where we really need to help people further is where their function has sadly got worse and they are entitled to more support, and they need to come forward. We need to be able to spot that. So that is the challenge. How do we get there?

A lot of it is around improving access to data. We are doing a lot of work. We have been working with colleagues in the Department of Health and Social Care on the relationship between information in the NHS and information in the benefits system. There is other work we can do to make it easier for people to share where their function has changed. And of course we need to get on and try to correct those situations where we have identified an overpayment. On disability benefits, I want to flag up that there is an underpayments issue just as much as there is an overpayments issue. I don’t know whether Bozena wants to say any more about what we are doing in that space.

Bozena Hillyer: You are right that underpayments is one thing we are very much concentrating on. The single biggest reason why there is fraud and error in personal independence payments and disability payments is functional need. There is no ready set of data and no one clear thing that we can go to that will help us in terms of the technology or the data analytics. This is what we are working on, much more in the error space than in the fraud space. It is about making sure that claimants have the right information so that they understand when their condition changes and when they need to tell us about it. It is about working really closely with the director for disability and PIP around how we can reduce the error that we see there and how we can help claimants get it right, because that is where the biggest issue is. It is absolutely about very much focusing on underpayments, because that is a real serious issue for those of our customers, particularly the most vulnerable, who are not getting the amount of money that they are entitled to. This is going to be a challenge because it is about information and understanding, and at which point they come and tell us that something has changed. I am not going to say that this will be an easy one to crack, but we absolutely do understand what we need to do here, and we are doing the work to try to improve that going forward.

Q33            Mr Mohindra: As a Committee, we are concerned. Obviously, people deserve the correct levels. The underpayments is a bit easier from our perspective. The worry that we have is the risk associated with trying to reclaim and the time it takes for an overpayment, especially if those people are not employed, and that might be covid-related. Are there any timeframes involved in recovery? The absolute backstop is deduction from state pension, but that could be years or decades away.

Bozena Hillyer: Recovery is dealt with through our debt teams in the way that every other type of recovery is, so there is no different timescale on that. We carefully consider whether there are cases of hardship, in which case we might recover at a much lower rate or we might give a pause.

We are very much on board with Breathing Space, which is the new initiative that will be starting next year to make sure that where we have vulnerable people, we are looking at whether they are actually able to pay back when we need to make a deduction. There is all of that going on as well. For those kinds of customers, there is not specifically any different timescale. We have the same timescales for collecting everything, but where we vary is on the particular circumstances of the individual.

We harness our people. We get staff referrals around this, but we also get public referrals. We have other sources of information, if not data, around where overpayments might be being made. Again, I do not want to say too much about it, but I have capacity and capability to investigate certainly the most egregious cases that we come across. It is always about that balance of dealing with those who are vulnerable in a compassionate way but dealing with those who are basically trying to take advantage of the system by going down the prosecution route.

Q34            Mr Mohindra: To be clear—this may not be your Department; it may be the debt recovery side—is there ever an opportunity or an instance where the debt is written off?

Bozena Hillyer: Actually, debt is my side of it as well, because I have all the teams that recover all the debt for the DWP. Yes, there are a few instances and cases where things are written off, but actually we have very good and wide-ranging powers in terms of debt recovery, so we can make attachments of earnings and we can take deductions directly out of benefit, all the way through to pensions and, indeed, taking money out of an estate at the end of life. There are cases where things are written off, but they are exceptional cases, I would say. Otherwise we look to use our powers to make sure that we get as much as we can back in for the Exchequer.

Q35            Mr Mohindra: We are obviously going through the covid pandemic at the moment, and we have seen a spike. Are we expecting—actually I might be encroaching on my colleague’s question later, so I will stick to my next question about housing and housing costs. Mr Schofield, why is it so difficult to verify housing costs?

Peter Schofield: The numbers on housing benefit per se have come down, but—sorry, I am slightly answering a different question, because that is helped by a better understanding about earnings coming from the RTI feed, something called Verify Earnings and Pensions. On housing costs per se, it is fair to say that we are making good progress on that in some respects. The landlord portal enables us to get access to housing costs for the social rented sector directly, which gives us a really robust approach to those costs.

In the private rented sector, we have typically relied on seeing the tenancy agreement. We are now automating that; we are rolling out an upload facility so that can be done remotely as well. It is an area where we are making progress. It is one of the areas we need to look at. I did not put it in my top 3, but maybe it is No. 4 in terms of things that we need to make good progress on.

The key thing—Bozena has talked a lot about the data matching—is building the relationship with landlords. We have done that for the social rented sector, and we need to build that further on. Part of that is in some of the features that Neil is working on. I do not know, Neil, whether you want to say any more about how we are building that data matching on the housing cost side.

Neil Couling: I think there are 1 million-plus private landlords in the United Kingdom, so an automated solution just doesn’t work. Most of them have one or maybe two properties—there are some bigger ones. The upload feature is for claimants to upload their tenancy agreements to us. It illustrates an interesting dilemma, which the Committee may want to think about: that will put time in the customer journey, because not everybody will remember to upload their tenancy agreement, so you will see a slight decline in payment timeliness as I roll out this feature across the country.

It is already at 90% or so—much higher than the legacy benefit system—so in my view there is a reasonable trade-off between getting greater accuracy and timeliness. In another part of the forest, I have the Work and Pensions Committee urging me to get payments to people even faster. These are what I call the trade-offs and the opportunity costs in trying to optimise a system design.

Q36            Mr Mohindra: Reverting back to the data—I appreciate that uploading tenancy agreements will help get a better understanding of the housing costs—is there an element of regret in not having local authority housing benefit fraud teams? Surely they would have had on-the-ground knowledge of the appropriate levels?

Peter Schofield: That has been a very effective way of driving out housing benefit fraud. We have seen some of that in the numbers on housing benefit. The main thing on housing benefit, as I say, has been linking that to a better understanding of earnings, but we have seen housing benefit fraud levels coming down.

The key is to find other ways of working, such as working with local authorities—I am keen for that to be part of what we try to do—and using the sort of automated systems that we have already described. Local authority landlords are very much covered by the landlord portal, so that is an automatic part of what we do already. Automating it so that we can do it in realtime is really the future. I come back to what I talked about earlier: our key priority is prevention, rather than relying on detection and collection further down the track.

Mr Mohindra: Chair, I should have declared prior to that question that I am a councillor.

Chair: You talk a lot about the private rented sector, and those 1 million properties that you mentioned, Mr Couling, are very varied in their management. This is an interesting area to probe—how you can have a relationship with what is essentially a bunch of private sector individuals and small businesses that aren’t really geared up to working in this way, rightly or wrongly.

Q37            Shabana Mahmood: I just want to explore with our witnesses fraud and error in the covid environment, and how we are going to unwind some of that. Mr Schofield, I completely understand that you have had to ease some of your controls over your systems in order to get money out the door very quickly to people in desperate need. Of course, that increases the scope for fraud and error. When will you know the total fraud and error cost of the control easements that you put in place as a result of the pandemic?

Peter Schofield: One of the things we are doing in realtime is, as we have had to put easements in place, we are also putting in place mitigations all the way through. We can go into more detail in private on some of that. These are things that we are working on in all the different elements, where maybe a face-to-face check is no longer possible.

One of the things that we have done in order to measure that is monitoring very closely the referrals that our staff are making back to our centralised enhanced checking service, and seeing how that is changing over time. Over a period of time, that peaked at one level at about 13%, but it has since fallen back to about 3% on average. Let me double-check that number. My memory was right: 13%, back down to about 3%. That is one of the ways that we have been monitoring that.

Obviously, Nick talked earlier about our publication of fraud and error numbers at the end. For me, as accounting officer, the key thing is to be able to monitor what we are doing in realtime. That is one of the ways we can do that. Our realtime analytics, about which we can go into more detail privately, also enables us to monitor that. The key thing is to have a real-time data feed that enables us to put our mitigations in place and to check that they are being effective.

Q38            Shabana Mahmood: I appreciate that, but the mitigations—I am glad you are doing them, and that you are not able to tell us about them, so that they can be as effective as they need to be in clamping down on fraud and error—will not eliminate the fraud and error. To be clear, I imagine the surge in claims will cause a lot of problems, so it is not a critique of your long-standing problems with fraud and error. We have already explored those in the hearing. I am just trying to establish that there is going to be a surge in further fraud and error as a result of the pandemic. I appreciate that you are doing mitigations to limit that amount, but you are obviously still expecting a very large increase.

Peter Schofield: There is certainly a risk, and we are looking carefully at that. As I said, one of the great guides here is the number of cases being referred to our checking services, which is now down to about 3% of claims—

Q39            Shabana Mahmood: That is staff referrals, but your staff are dealing with a level of claims that they have never had to before, and you have a lot of new staff and people from other Departments, so with the best will in the world, it is not physically possible for your staff referral system completely to catch all the cases where they think there will be a problem. I think we can all accept that there will be a big issue to resolve once we are through the pandemic. I appreciate that you want to say that you are making progress, but there is perhaps a degree of over-optimism creeping in about the scale of the issues you will have to unwind once we are through the pandemic.

Peter Schofield: I am not complacent, but I would say this to the Committee: we are assessing this at every stage. We have put easements in place and then we have switched them off, if I may put it that way, and even on the face-to-face side of things, I have 640 jobcentres and they are now all covid-safe and able to invite people in for face-to-face meetings, if we need to do that. Even with that element, there is an end to the easement, if I may put it that way, on top of the progress we are making with predictive analytics and data matching, which are automating processes that have always been manual in the past.

At the end of all of this, I am quite sure that there will be cases—you might also push me on this—that we look back on and say, “Okay, these cases are in payment and we need now to look back at some of the ones where there might have been a higher risk around them.” This again is where data analytics comes in to help us.

Without going into a huge amount of detail on this, we have categorised cases according to the nature of the case, the composition of the case, and the way that it came through the system. Bozena might correct me on this, but I think we ended up with something like 84 different categories. We are now going through trying to assess the risk around those categories and then deploying staff. We are doing this on a test and learn basis at the moment with 100 staff, going through the highest-risk categories to assess how we would go back and—as we put it—repair those claims.

That is trying to identify fraud that has got through the system despite our mitigations and all the work we are doing, so we are going back through those cases. As we develop that, and the early signs are promising—I will hand over to Bozena—that will enable us to claw back those cases that got through despite our mitigations but where we can still stop payments in the future. Bozena, do you want to say more about that process?

Bozena Hillyer: As well as staff referrals, which is fantastic, because our people are brilliant at spotting stuff and want to tell us, we also put in place what we call the integrated risk and intelligence service. They were all elements that were there and were coming together, but we placed them in one virtual team. I will not say too much about the type of capability that we have, but we are able to look at what is coming into the system, along with other types of intelligence and data.

We put that in place by the end of March. People like to call it the war room, which is probably going a little far, but there is a daily war room to look at what has come in and what we are seeing. The data analysis we have been able to do through that has informed this identification of what we believe are the riskiest cohorts that have come in. As Peter said, we have already started doing some test and learn investigations on what we think are the riskiest cases. They are indeed proving to be the riskiest cases, so the fact that we are identifying the right ones is a great tick in the box for our risk engine, and we will be doing that test and learn over the next few months. Out of that will come a plan in the new year for how we then go back and make sure we repair everything that came through when we had easements in place.

I would also just emphasise that because something came through when the easements were in place does not mean it is wrong. The vast majority will be right and will be fine; it is about finding the riskiest needles in that particular haystack, as you referred to the volume. I think what we have done so far, and the results we are getting, are showing that we have identified where the risk is.

Q40            Shabana Mahmood: Thank you, both of you, for that level of detail about not just the easements, but the mitigations and the work you are doing with your new systems. All of this realtime data analytics sounds fantastic and very positive, but that rather goes back to my very first question: when, therefore, will you be able to tell us and Parliament the amount of fraud and error cost from the easements? You have all of the data, and everything you have told me sounds fantastic, so I am wondering why you are not in a position to tell me at which point in time after we are through the worst of the pandemic you will be able to report accurately to Parliament the cost in terms of fraud and error of the easements that you were forced to make in order to deal with the pandemic, so that we can hold you to account.

Peter Schofield: We will aim to do that in our annual report and accounts. That is the point at which the analysis will have been done—the sampling that Nick and Neil have already talked about. We will then be able to set out where we were in terms of fraud and error over the course of this year, 2020-21. We will look at how much detail we can go into about attributing that.

Q41            Shabana Mahmood: Given the detail I have heard, Mr Schofield—this is a warning shot for a future hearing when you come to talk to us about your annual accounts—your explanation for a deficiency in your knowledge had better be that you did not have enough data, because everything you have told me today rather implies that you have got lots of data. You have all these new, whizzy systems, and I really hope they are in place by the time we have a hearing about your annual accounts.

Peter Schofield: That is a very fair challenge, Ms Mahmood. I look forward to it.

Chair: I bet you do.

Q42            Shabana Mahmood: So do I, Mr Schofield. Have all of the controls that you had to ease been reintroduced? I appreciate there is a difference between those for whom you put in mitigations, but are we now in a position where all easements have been phased out?

Peter Schofield: The main one that we still have in place is that it is not standard practice at the moment to invite new claimants in for an initial evidence interview in jobcentres.

Q43            Shabana Mahmood: But everything else, apart from that? Ms Hillyer is nodding.

Peter Schofield: That is right. I was just racking my brains as to whether there is anything else, but that would be the main thing.

Q44            Chair: So basically, the only thing you have stopped is the volume physical contact bit of it.

Peter Schofield: Neil may want to add something. He may have thought of one that I forgot.

Neil Couling: Easements is quite a broad description. If they are easements relating to our controls on the system, Peter’s answer is quite right. There are other things that we are still not doing: we are not doing full conditionality on every case, for example. I suspect that will have some sort of fraud and error impact down the line by the time we get to the accounts; the question is whether we can identify it. The brilliant systems they have been describing to you will not bring that one out, I am afraid, but in general, we have prioritised putting our controls back as fast as we could. We were trying to unblock claims because of the health emergency. We did not want people without money, unable to isolate in their homes, which is what the early part of lockdown was all about.

Peter Schofield: I have just thought of another one. We are largely doing health assessments over the phone, rather than face to face, so that is another area.

Q45            Shabana Mahmood: All of which I guess will continue while we are still following public health regulations because of the pandemic. That is helpful—thank you.

How are you using the unwinding of the easements that you can proceed with at this point to help you understand the cost-effectiveness of your controls? What insights do you have as to whether those controls were any good anyway, or what impact they have on your overall numbers?

Peter Schofield: That is a good question. It plays into the way that we use the realtime information that we have, both from our predictive analytics, which we talked about earlier, and the staff referrals. We are able to see, as we unwind things, whether they make a difference. I am not sure that I wanted to talk about any individual one, actually. There was one that I was going to talk about, but I am not sure that I will publicly. We are able to look at what the impact is and whether it has stopped that fraud or error coming into the system because we see, as we switch it on, what difference it has made.

The other important thing is that this covid situation has actually accelerated a lot of new ways of doing things, which are proving to be more effective maybe than what we were doing before. One that I will talk about is building on Verify, with our own Confirm Your Identity system. You may remember from a previous session on universal credit, with Verify only about 25% of claimants were able to successfully verify through that cross-Government system. With Confirm Your Identity, we have been able to push the total number of online verifications up to 40%, and we think that we can go further. We probably would not have accelerated Confirm Your Identity’s rollout were it not for the pressure to do so because we were not doing the initial evidence interview in jobcentres.

Q46            Shabana Mahmood: That is a good point about accelerating new systems and best practice—I am glad that is happening. I appreciate that and do not want you to give details of any particular controls, but just so we have a sense of numbers, you have confirmed that there is one control you had that did not seem to add very much and which you have removed. What number of controls are in the same position?

Peter Schofield: Sorry, if you took from my previous answer that the things we were doing before were not effective, that is not really what I meant. I meant that there are better ways of doing things, particularly with some of the automations. There are some big areas where we have—I do not want to go into a lot of detail publicly, to be honest—

Q47            Shabana Mahmood: Just give me areas or numbers that you can tell us about.

Peter Schofield: Bozena, do you want to have a go?

Bozena Hillyer: This is difficult without talking about something specific, but there is an area in which we were seeing an increase. I talked about using models and risk engines to try to identify how many are high risk in a particular benefit or in a particular type of claim coming through. To give you a sense of scale—although I cannot talk about how we did it or, necessarily, the thing—at the peak, we were detecting, and our model was showing, that around 7.5% would be in the high-risk category. We then put back certain controls and that has now dropped right down to 0.7% now being in the high-risk category, so we do have information and understanding about which of our controls actually has an effect. Apologies if that sounds a little bit vague, but it is because I do not want to give away exactly what it is that we are doing and how we are doing it, but we definitely have the numbers to show that it has had a very good impact.

Q48            Shabana Mahmood: That is helpful. May I ask you about the recovery of overpayments? During the pandemic, what are you planning for in terms of the potential numbers of prosecutions or an increase in prosecuting activity?

Peter Schofield: Bozena will probably want to come in on that. One thing we have been discovering, when you get to that stage in the process of recovery, is the most effective balance between prosecutions, on the one hand, and more administrative processes, such as simply writing a letter asking for the money, which is actually generally quite an effective way of getting some low-level responses back. Bozena, do you want to talk a little bit about what we can say on that?

Bozena Hillyer: You are absolutely right. It goes back to the very helpful National Audit steers around value for money and cost-effectiveness. Prosecution is obviously an expensive way of deterring people. We will always do a level of prosecutions, partly because there is some very bad behaviour out there; there are some bad people, sadly, who look at the country coming together over covid and see an opportunity for getting money. Those types of people we will always want to prosecute, and we will do that to the best of our ability.

On recovering any overpayments that have come out through covid, our No. 1 priority is to try to get as much of that money back as possible, and we are looking at that really carefully, using the lens of behaviour, which is a learning from other Departments that I brought across from HMRC. Where you have people who are at the very mild end—it might be a mistake, or something that is not too bad—we are going to use letters or other ways of asking them to pay the money back. We will reserve prosecutions for those cases where it will either provide a great deterrent effect, because it will be an example that we can use in public, or because it is the very worst behaviour that we see.

Obviously, we have had to suspend interviews under caution through the covid period, because of the social distancing rules, but we are looking to stand those back up later this year. I am actually in the process of recruiting extra people into my serious organised crime and economic crime group teams, in order to make sure that we are ready to carry on doing the disruption work, which we have found incredibly valuable, and also to look at prosecuting in the instances that I have outlined where we will want to do that.

Q49            Shabana Mahmood: I am thinking of HMRC in particular. In this pandemic phase, there are obviously people trying to defraud the benefit system, and also all the other loans that have been made available for businesses that are experiencing difficulties because of the pandemic. From what I see, just in my own city, there is clearly organised crime activity on both sides of that equation—on benefits, as well as on bounce back loans and so on. Yourselves and HMRC will both be picking up an increased risk of fraud but also the possibility of knocking out some fairly organised criminals at the same time. What work is going on across government to make sure that you use your resources effectively and prosecute in an organised way to try to deal with this threat?

Bozena Hillyer: We always work, and have always worked, together with the National Crime Agency, the police, HMRC, the Health and Safety Executive, the Gangmasters Licensing Authority, and so on. Through covid, we have come together even more closely, under a range of new groups, kind of sponsored by the Cabinet Office, and through things like the Joint Fraud Taskforce, to make sure that we share intelligence and data.

We are well used to working very closely with the police and others on prosecutions. As you say, if somebody is acting in a criminal way in one place, they are probably doing it in another place as well. We very much look to work together, and around things like the detection of modern slavery and human trafficking, we work together very closely to bring joint and overall prosecutions under different pieces of legislation. That is there and it will continue.

Q50            Chair: You have talked about suspending the under-caution interviews, but social distancing rules would allow those to continue with the right safeguards in place. Why have you dragged your feet on getting those back if they are so important?

Bozena Hillyer: We haven’t dragged our feet. We have to, first of all, make sure that our own people and customers are safe. Partly, this is just a physical problem. A lot of the estate that we have does not actually allow us to do that in a socially distanced and safe way, so what we are doing at the moment is finding other solutions and looking for other ways in which we can do it in a socially distanced way. It is purely a question of finding a way that we can do it to make sure that both our people and the customers are safe.

Chair: We get that that is important. It is just so important to set a mark on this.

Q51            Shabana Mahmood: We have just seen the announcement from the Chancellor that the Government are going to pay up to two thirds of wages for people up to six months. Obviously that is going to impact very much on your workload, Mr Schofield, so were you aware in advance? Also, how are you geared up for dealing with variable income and having to calculate it at this two-thirds rate?

Peter Schofield: We were consulted, but obviously not in a huge amount of detail, and I have not watched the statement because I have been with you, which has been nice. Neil will want to come in on the question about variable income, but of course universal credit is absolutely set up to be able to manage around variable income. That is also measured through RTI. One of the benefits of the system that we have is that it is resilient to this, so we will look and see how this is taken forward, and I look forward to reading in detail what the Chancellor has said.

Neil Couling: The key thing in the discussions with the Treasury was to reassure ourselves that this would be paid through PAYE, which it is. That means that it links through into RTI, which means that UC will work in the normal way. Clearly, there may be a change in the level of earnings because of this, because of the third third third approach that the Chancellor is taking in the successor scheme to furlough, but that will all run nicely through the RTI system for us, so it does not create any additional burdens on us. It probably helps us in terms of keeping the level of unemployment down, so we will not have as big an increase as some of us were fearing. I have been watching the skies nervously for weeks now.

Q52            Shabana Mahmood: That sounds very positive. I hope that that positivity is reflected in the experience that we have. Obviously, RTI is problematic for some cohorts of applicants. Is it your position that what has been announced, and what was discussed with you prior to the announcement, is easy enough to administer from a system perspective, but that the people who have struggled since the pandemic began and since we had the first furlough scheme will continue to be the same types of cohorts of people who have struggled—freelancers, the self-employed and so on?

Neil Couling: RTI works very well. We did over 20 million calculations through RTI last year, and there were only, I think, 20,000 challenged as wrong, and that was employer failure from putting the wrong RTI in. We are dependent on employers. The benefits system is there as a safety net for people not covered by these schemes.

Q53            Shabana Mahmood: I am getting messages through; I am stuck in this hearing with you as well, so I am unable to see the Chancellor’s statement. The furlough scheme has been subject to fraud. Will the fraud from the furlough scheme flow into universal credit?

Peter Schofield: I do not see how. I am just trying to think. What lies behind the question? Is there a particular thing that you are concerned about, or is it just a general worry?

Q54            Shabana Mahmood: There is clearly fraud within the furlough scheme. As that is changed and we move to this second phase of furlough, essentially, obviously there is a concern about the fraud flowing into your systems because you are now administrating the next phase. I am just wondering what your concerns are.

Neil Couling: Let me have a go here. If I apply under the furlough scheme as a fraudster, I get money that I should not be entitled to, but that shows up in my PAYE, which means that any universal credit will be lower. That is a fraud issue to manage, and my colleagues in HMRC will be trying to manage that, but in terms of the effect on us, they are showing us that they have a higher income because we get that PAYE feed. I do not see it necessarily following through in that way.

Q55            Shabana Mahmood: So it should be picked up by HMRC one stage earlier.

Neil Couling: We know from Jim Harra’s evidence that Jim is all over trying to manage and keep the fraud in those schemes as low as he possibly can, much in the same way that we are trying to do the same in the DWP benefits.

Q56            Chair: We understand that the Chancellor has made some announcements about self-employment. Do any of you know anything about that and how that might impact on those already tricky claims?

Neil Couling: I am aware that there were thoughts around self-employment. Again, because of the way in which universal credit works, payments under the self-employed income support scheme were taken into account. I do not know exactly what he said today, but my understanding is that, if they follow the same pattern through, universal credit will be perfectly capable of coping with that new scheme.

Chair: Okay. You are very upbeat about the ability of universal credit to cope, so we will hold you to that next time.

I thank our witnesses very much indeed for their time. Your report and accounts are a hugely important annual event for us. We are heartened, Mr Schofield, that you are keen to set targets for individual benefits and more generally on fraud and error, and we will be holding you to that and to deadlines and timelines. Thank you very much indeed.