Reducing errors in the benefits system - Public Accounts Committee Contents


Examination of Witnesses (Questions 1-158)

Q1 Chair: Welcome to you all, and a particular welcome to you, Sir Leigh. I think this is probably your last appearance—out of how many?

Sir Leigh Lewis: It's hard to get this right. Neither the Committee Secretary nor I are able to work this out, but by my very crude reckoning, we're up to about 30.

Chair: Thirty. Well done.

Mr Bacon: If you get it wrong, just put it down to administrative error.

Mrs McGuire: You said we were going to be nice today.

Q2 Chair: This is quite a topical session, given the judgment that I heard on the one o'clock news on the radio. You might be able to tell us a little bit more about it. It is that if there is an error that is caused by your administration, you will not in any way be able to recoup it. This makes this issue, as I understand it, ever more important, because once you've made the error, you have to write it off. We've all looked at this, and I know it's hellishly difficult; we can all understand that you're dealing with a huge number of individuals, a lot of claims and a lot of members of staff.

If I can ask you, Sir Leigh, to reflect on it, you've been trying to get administrative error down for ever, and yet you're finding it very difficult to shift. Whilst as a percentage of the total money that you give out it's not enormously high, in absolute terms it's a lot of money to the taxpayer and to Government. What's gone wrong? Why have you not been able to get it down better than you have? What's gone wrong?

Sir Leigh Lewis: Thank you, Chair. Just to say, as I depart, it will be with one clear failure. I arrived just over five years ago back into the Department, and on my very first day I remember saying to all the Department's senior staff that I wanted to achieve the lifting of the qualification of our accounts in that period, which had then been there for 16 years. I haven't achieved that; as I go, it's now been there for 21 years. We might come back to that at some point.

Why is this so challenging? Two sets of reasons, if you'll let me talk about them. First of all, you can always do better. We have put in enormous effort, and enormous energy and commitment—the Report from the National Audit Office acknowledges that, and I'm grateful—but there's always more that you can do, and more systematically and better. The Report, I think, points us in some directions where we can do better still, and do more, and use information and data. This is not a story of, "We are doing everything we can in the best of all possible worlds."

I will put that to one side, and reflect a little bit on a benefits system that is unbelievably complex. At the risk of losing any support from the Committee before I've even started, in a way we're all responsible: successive Governments and Parliaments, as well as successive administrations and civil servants. Depending on your definition, and not counting passported benefits, there are something like 30 separate benefits in our benefit system. It's interesting that the Report says there are "over 27"; we can't quite get to how many there are.

In terms of possible combinations of benefits, it's simply impossible to arrive at a definitive figure. They certainly run into hundreds. We've got 25 special groups of people in our benefit system, with their own sets of rules: war pensioners, crofters, seasonal workers, share fishermen. We have 900 distinct rates of payment. We issue 14 manuals of 8,690 pages to our decision­makers.

Going back to 2006, the PAC, in its Report Tackling the Complexity of the Benefit System, said: "It is hard to quantify how much the system has grown, but the authoritative Child Poverty Action Group guide to welfare benefits… has grown from 432 pages in 1990-91 to 1,546 pages in 2004-05." It has now reached 1,680 pages. Maybe it's always a good thing to quote one's own Secretary of State. Ian Duncan Smith said, on 11 November when he launched the White Paper on Universal Credit, "Our antiquated welfare system has become a complicated and inflexible mess." In a sense, that's the system that we're trying to administer.

One other thing, although I really don't want to bore the Committee. Society and societal change in that period has become ever more complex. We live in a society in which people's lives change at a rate that my parents, let alone my grandparents, would have found unimaginable in terms of the complexity of people's lives. If you look, for example, at rates of divorce, they've gone staggeringly higher since the modern welfare system came into existence in 1948, let alone since I entered the Civil Service.

If you just take one factor, one figure, which I think is quite staggering, births outside marriage—this isn't a moral statement of any kind; society simply is the society that it is—were at 5.4% in 1948, when the modern welfare system was introduced. In 2008, they were at 45.3%. Not only have we got a hugely complex system, but we're trying to keep track of people's very fluid and ever­changing lives in that system. We could do better, but in many ways, I think against a background of a system of that unbelievable complexity, perhaps it's not altogether a bad figure that we lose 0.7p in the pound through official error.

Q3 Chair: I think all of us accept that. You could put the other side of that equation—for example, that advances in information technology mean you can communicate much more quickly than you might have done 30 or 40 years ago. That ought to allow you to have much better standardised ways of administering what is undoubtedly a complex system. The equation is two­sided, and we have to understand that.

What I felt, looking at it, if you look at the system as a whole, yes, it's complex, and yes, everybody's added to it every time there's been a new administration. But there are key areas where you get a concentration of administrative error. Homing in on those, which is what this Report does, you might have expected to have greater success, and yet, if you look at the figures, they're going in the wrong direction.

I know we've had a recession, and I know that's resulted in more JSA applications. Nevertheless, if you look at income support, if you look at all those means­tested areas where you will get a concentration of error, the figures are going the wrong way, not the right way. It just seems to me that if one had had the focus on those few areas where you had the greatest problems—or indeed on the people who had the greatest problems; the people who have had over £10,000 more than they should have done, out of administrative error—you might have got further. I can't quite understand why, in all the work that we know you've done, you haven't been able to drill down to where the problems really are.

Sir Leigh Lewis: Yes. A frustration for us, as well as a frustration for you. I think the Report, in fairness, points to some ways and some techniques where we could do it better still. I do understand and take that. It's been a very useful exercise. But it's not as if we have not been doing this; it really is not.

I don't want to pretend that nothing was being done before I arrived, but I did put a huge personal emphasis behind this, and that's been there throughout my time. Over the last five years, we have focused in on the most error­prone benefits; we haven't spent a huge amount of time looking at state pension, for example, where the official error rate is 0.1%. We've homed in on those benefits that have had the highest rates of error. Within that, we've homed in on the top 10 causes of error. We can go into all of these areas in more detail. We've looked at where we can use IT—your point exactly, and fairly—to try and take error, design error, out of the system.

In fairness, the overall rate of official error, as the Report says, has stayed pretty constant at about 0.7% through that period. It's gone up 0.1%, or it's gone down 0.1%; it's barely changed. Incapacity benefit has got worse. Some of the other benefits have got seriously better. In pension credit, we're at our lowest level ever of official error.

But it's been not for want of trying. That is one of the advantages of moving towards a universal credit, which will be, if it's implemented as the Government would wish, a radically transformed and simpler form of benefit for people of working age. I don't think, with the best will in the world, simply going on as we have is going to fundamentally bring this down further. It will in some respects. We can do better, but if we're really going to crack this, we have to have a radically different design of the benefits system.

Q4 Chair: The other thing that I would ask you is about local government, which appears to be a particularly bad culprit on this. Somewhere in the Report it says that local government "accounts for 17% of overall benefit expenditure and for just under 25% of the overall level of administrative error." Have you got a comment on that aspect? That hits you, and you think, "Ah, is there something there that you could do?"

Sir Leigh Lewis: Yes. It's true, as you say: housing and council tax benefits together account for about 17% of overall expenditure, but for about 25% of the overall level of administrative error. It would be comforting, in a sense, because we don't do that, for me to sit here and say, "Oh, I know. For goodness' sake, why don't they pull their finger out?" In fairness to many local authorities and the work of my own Department, an awful lot of effort has gone into that. I think one of the issues there is that, with well over 300 local authorities administering housing benefit, you get into the issue of performance variation, inevitably.

Q5 Chair: Would you centralise it?

Sir Leigh Lewis: I would consider that strongly, and that is part of universal credit. Yes, that's part of the design for universal credit.

Q6 Chair: It's got a history. It used to be administered centrally, and then, I can't remember when, it was decentralised to local authorities.

Sir Leigh Lewis: Yes. I'm not sure I even know the history. I think if you were starting with a very clean sheet of paper, saying that you want essentially one benefit to be administered by over 300 different bodies, probably you'd just pause at that point, before you ticked it.

Q7 Chair: Yes. I remember the history. It was because it was so inefficiently done from the centre. It wasn't done for ideological reasons, but it was an efficiency argument—that you might have fewer errors if you did it at the local authority level than if you did it at the central Government level.

Sir Leigh Lewis: It's certainly worth saying that the design for universal credit, which is, of course, a very radical transformation, would effectively bring that within a single working area

Q8 Chair: That's 10 years down the line.

Sir Leigh Lewis: Not quite so far.

Q9 Chair: Eight years.

Sir Leigh Lewis: Starting in October 2013, according to the Government's plans.

Q10 Chair: As these things go, probably 10 years. Richard says it's probably longer. Were you to remain a Permanent Secretary, you would say, "Centralise housing benefits," would you?

Sir Leigh Lewis: I would say, "Don't do it at the drop of a hat, because that would be the most enormous task." Perhaps you would expect me to say this, Chair, but I do welcome the plan for a universal credit. It does seem to me that it is the radical re­engineering of the benefits system that is required.

Q11 Stephen Barclay: Just picking up on the point that the Chair is making—that perhaps a solution is a further eight years away—can I just take you back to the PAC Report in 2002-03, Tackling Benefit Fraud? If I can quote its findings, "The Department's experience over the years suggests that the arrangements whereby local authorities pay housing benefit are insufficiently robust to deliver the benefit without substantial losses to the taxpayer. If the Department cannot make these arrangements work more effectively"—as you've just alluded to—"they will need to consider what alternative methods of delivery might align payments better with lawful entitlement."

If this Committee was saying exactly the point you've just made, and they were saying it back in 2002, and the solution is a further eight years away, why is it taking so long?

Sir Leigh Lewis: First of all, just to get our timings right, the Government's intentions are that universal credits should start for new customers to benefit in October 2013, and then it would progressively take on the existing stock over a period. This is a huge challenge. It's certainly not tomorrow, but it's not as far away as that.

Two things, in answer to your perfectly fair question. First of all, no Government, in that intervening period, has decided to fundamentally restructure the housing benefits system, because while that Report may have been right in its time, this would always have been seen as a massive and major step. As I know, re­engineering the benefits system is not something you take on lightly.

The other thing I'd like to say is throughout that period, both local authorities and my Department have worked very hard and systematically to improve the administration of housing benefit through local authorities. There are some local authorities that are absolutely first­class on this, and we do operate now a whole system of carrots and sticks with local authorities to encourage better performance.

Q12 Chair: Has the level of administrative error gone down since 2003?

Sir Leigh Lewis: I think overall the level of administrative error, if you take housing benefit fraud and error together, is now at a record low of 4.4%. That's down from the peak of 5.5% in 2005-06. It's down, but it's still at a high level.

Q13 Stephen Barclay: Could I just come back on those figures, because I looked at your Annual Report for the first year, going back for five years, given that you've been in the post five years.

In terms of fraud relating to housing benefit and council tax, in 2005-06 it was 1.2%. In 2009-10, it's 1.2%. It doesn't seem to have moved. The quantum in 2009-10 is £300 million. We picked on the earlier PAC finding that said perhaps there should be a fundamental shake-up. That hasn't happened. This Report says that, even without that, you're not using your resources effectively. That's the finding on page 9, where it says, "We could not establish that the Department had consistently targeted its resources and initiatives to the best effect."

When there was the report Progress in Tackling Benefit Fraud in July 2008, the Department reported that it was spending £154 million combating fraud, for the year 2006-07, but it only recovered £22 million for that spend. Can you just clarify, in terms of the figures here, if there's a fraud level for these two benefits of £300 million, how much is the Department spending investigating that, and how much is it recovering?

Sir Leigh Lewis: There were quite a lot of questions in there, so let me try and just take them, and if I have missed one along the way, then you'll come back to me.

Stephen Barclay: Sure.

Sir Leigh Lewis: First of all, I think the figures that you're quoting were the figures, understandably, within the report for official error within housing benefit. That latest figure is at 1% in 2009-10.

Q14 Stephen Barclay: No. Sorry to interrupt, but it's also £300 million. You're absolutely correct it's the same in 2009-10. If one looks at the official error figure for 2009-10, coincidentally—and this may be just because the Department is so cavalier with figures that it rounds them up to a neat number—it's £0.3 billion, 1.2%, on official error. It is also £0.3 billion for 2009-10 on fraud.

Sir Leigh Lewis: First of all, one thing I have learnt, in 30 appearances before the PAC, is not to accept throwaway remarks that aren't true. The Department is absolutely not cavalier on its measurement of error. Indeed, it's been complimented by the NAO, in its Benchmarking Report, as having a better record of measuring fraud and error than any comparable administration that the NAO were able to identify.

Let's return to a point of agreement: it is still too high. The figures I was quoting—and they're entirely compatible with the ones you're quoting—were the overall total for fraud and error in housing benefit and council tax benefit as a whole.

Regarding two other things that you said, and let me try to pick those up—cost-benefit, and is what we're doing cost­effective? As I think figure 8 of the Report shows, for the various initiatives that we undertake—and this goes wider than housing benefit, in fairness to your question—we believe that we save between £3 and £70 for each pound we spend. Overall, we believe that the error reduction strategy that the then Government published in 2007 has saved over £800 million since its inception.

The other point you raised was debt recovery: how much do we recover from the debts, whether they're from fraud or error? I haven't got those figures in my head, but they've been going very substantially upwards. Indeed, this Committee and the NAO produced a Report, in the last couple of years, on debt recovery. Our debt recovery is at record levels, and we recover getting on for £4 of debt for every £1 we spend. None of these things are ever good enough, but they are all ones to which we have devoted very considerable effort.

Q15 Stephen Barclay: With respect, that's a different answer from the question that I put. The question I put was more targeted: it was about housing benefit and council tax benefit, because what concerns me is that they are administered by local authorities. The question is: how are those local authorities incentivised? If you are paying the bills, how are they incentivised, particularly in tough times, to address fraud?

Sir Leigh Lewis: Okay.

Q16 Stephen Barclay: To what extent, for example, are they liaising with SOCA, as local authorities, in order to capture the data that SOCA has vis-à-vis some of the things it has on payments into accounts from banks, which relates to benefits fraud?

Sir Leigh Lewis: We're straying a little, of course, from today's hearing, which is about official administrative error.

Q17 Stephen Barclay: Some 25% of admin error relates to this issue.

Sir Leigh Lewis: I'm speaking essentially from memory rather than having gone through this in every detail.

We work very, very closely with local authorities on fraud. There have been some major advances—for example, we can now prosecute each other's cases of fraud. There is closer working between the Department's fraud investigation service and local authority fraud teams than there has been for a long time.

Something that I'm proud of having introduced in the last five years is a Back to the Floor programme, where senior members of staff spend a week each year working directly in one of our customer­facing parts of the Department. I spent a week myself, in August of this year, working with our fraud investigation team in the South West. As I saw for myself, there is very close co-operation in working with the local authority fraud teams. Fraud debt is hard to recover, by definition. You're dealing with people who are setting out to defraud you. They've not made an error by accident.

Q18 Stephen Barclay: Could we perhaps have a note?

Sir Leigh Lewis: Yes.

Q19 Stephen Barclay: We can't just rule out the fact that 25% of administrative errors relate to that administered by local authorities. What I'm trying to get is a figure for how much is being spent investigating that £300 million. Why is it that five years ago it was 1.2%, and this year it's 1.2%, so it's not moving as a percentage, and as a quantum it's going up?

I don't get a clear sense, with respect—this is what the NAO Report says is not happening—of the extent to which there's a cost-benefit analysis within the Department that the £27 million spent on adverts over the last five years to tackle fraud is more effective than using that £27 million to incentivise local authorities or using it in other ways. What I'm struggling to understand is how you're targeting your resource to address that 1.2% of fraud.

Sir Leigh Lewis: Let me let you have a note, because that would be a very good idea. Of course this Committee did have, and I responded to, a Report on fraud in the benefit system about—I'm looking at the Comptroller and Auditor General—I think two to three years ago. At that point, fraud was at 0.6%. It hasn't changed very much; it's now 0.7% of the total. This isn't just local authority, by the way; it's right across the benefit system. The conclusion of the Report was that we were probably close to an irreducible minimum at that level. Let me do you a note.

Q20 Ian Swales: I'd just like to return, because you mentioned figure 8, Sir Leigh, where you're looking at the return for money spent on the various interventions that you've carried out. Any businessman would look at that and say: "That's an excellent return." You're getting, as you rightly said, £5 to £70 return for every £1 you spend on interventions.

My question is: how do you know you're spending the right amount? Surely spending more is going to lead to a greater return, isn't it? At what point do you stop spending on interventions? To what extent are you limited by, you could say, artificial limits on how much you can spend? If this is the kind of return, then it's obviously the right thing to do this work for the taxpayer.

Sir Leigh Lewis: Absolutely. Two things we're trying to do, in that respect. One is that we're trying to use our existing resource to get the best possible return, in exactly the way. You look at those tables, and you say: "The one that's returning you £70 for each £1, that seems a particularly good use of resources, so let's see just how far we can push that."

In fairness, that's data matching. We've hugely expanded our data matching in recent years, precisely because it gets such a good rate of return. The Fraud and Error Council, which John Oliver, who's our Director of Benefits Performance, chairs, which reports directly to me, is absolutely tasked by me for looking the whole time at where we get the best return for our pound.

In fairness to successive Governments—this is not a party point in any way—successive Governments have been willing, particularly in recent years, to invest more money in order to get a better rate of return. Over the period from the year we're in, 2010-11, up to 2015-16, we will be investing over £450 million of extra resource in order to deliver additional fraud and error savings of over £2 billion. That's come through in successive Spending Rounds, PBRs, Budgets and so on. I could take you through the various initiatives.

Q21 Chair: Could I just ask you on that, given today's judgment, will you be focusing that more on administrative error or on fraud? Once you've done administrative error, you're not going to get it back.

Sir Leigh Lewis: Today's judgment, Chair, of course, we've only had today, so we do need the time to read it and study its implications, etc. I don't really want to say, "No, absolutely, we're going for that." What that judgment suggests is that the importance of trying to stop that error from getting into the system in the first place is ever more important.

Q22 Ian Swales: Just a final question on that point. Clearly, you are saying you've made a business case, if you like, to get this money to spend. What are the constraints? Would you like to have more? I'm sure you'll say you'd like to have more, but what's the business case for the taxpayer? If they spend another £100 million, how much extra will you get back?

Sir Leigh Lewis: There are always a number of factors, aren't there? First of all, somewhere there's always a law of diminishing returns; at some point, the extra pound stops buying you three pounds, and buys you one pound and tuppence, and so on. Secondly, successive Governments may see the intellectual case for that extra investment, but nevertheless it has to be provided out of taxpayer resources, so there are always expenditure constraints. The third issue is simply the capability of the system to gear up and do more. One of my pieces of experience is that simply throwing money at something, in a cavalier way, not planned and not thought through, tends to get you pretty poor rates of return.

Q23 Ian Swales: You think you've got about the right resources, given all those constraints? You think the situation going forward is about the optimum? Is that what you're saying?

Sir Leigh Lewis: I think it's pretty good, yes. Having an extra £450 million over the next five years—that's not trivial money at all—puts us in a good place. Even if somebody suddenly said, "Just help yourself," I wouldn't go and pick hundreds more millions out of the coffers.

Q24 Mr Bacon: Very quickly, Sir Leigh, when you say, "One of my experiences is that throwing money at problems doesn't solve them," can you give an example of where you've thrown money at something?

Sir Leigh Lewis: No.

Q25 Mr Bacon: You've just said it was from experience. Which experience are you referring to?

Sir Leigh Lewis: So I must have some experience, mustn't I? I don't think I'm going to get drawn into going back over inglorious episodes in my career.

Chair: You might provide us with a note on that.

Q26 Mrs McGuire: Thanks very much. To a certain extent, this dovetails a little into Stephen's line of questioning. Why do we put fraud and error together?

Sir Leigh Lewis: That's a very good question.

Q27 Mrs McGuire: That's why I asked it.

Sir Leigh Lewis: One reason is that it's tempting to think that they sit in completely watertight compartments, and that something is always the one or always the other. The truth is that there are lots of shades of grey. You know the extremes. The extreme on fraud is someone who sets out, very deliberately—on a massive scale, often—to defraud the taxpayer and the benefits system.

Q28 Mrs McGuire: So you could put them in a box?

Sir Leigh Lewis: At the other end, you've got customer error, where the customer, completely innocently and totally unknowingly, does something, tells us something, fails to tell us something that they should tell us.

The truth is that although we categorise, and the NAO when they audit our accounts categorise, there are lots of shades of grey in individual cases, where it's really quite hard to know whether there was a kind of malevolent intent or whether this was truly innocent. The other thing I'd say is that lots of the solutions, like data matching and the use of better IT and training, target both fraud and error. To put them completely in separate buckets or silos and treat them separately would be wrong.

Q29 Mrs McGuire: Do you think it's helpful to the debate about how we deal with fraud in the system to continue to talk about fraud and error as though they were totally interrelated?

Sir Leigh Lewis: You lead me into almost how these things could be reported and presented, and almost into political territory. There is a big difference between somebody who sets out utterly deliberately to wilfully defraud the taxpayer and the benefit system, and someone who is, at the other extreme, simply confused, baffled and bemused by the system, and makes mistakes because they can't understand it.

In a sense, I suppose, the unifying feature of those is that this is, in the end, loss to the taxpayer, and it's loss to the taxpayer in the benefit system. In terms of trying to tackle it, it is logical to put them together.

Q30 Mrs McGuire: We have big figures here in the Report. I have tried—I may have missed it—to find out what the average administrative error is costing. I'm not explaining this very well. The individual cost, or the cost that the individual is overpaid, that's what I'm trying to get down to.

Sir Leigh Lewis: Yes.

Ian Swales: Per person.

Sir Leigh Lewis: Yes.

Q31 Mrs McGuire: We've got millions of figures. The per-person average, in terms of administrative error. It may be in the report. Is it?

Chair: There is some.

Sir Leigh Lewis: There is some information on that.

Phil Gibby: It is very difficult, but figure 4, on page 16, gives you a breakdown by value for the three main areas.

Q32 Mrs McGuire: On an individual basis? Or an average?

Phil Gibby: It gives you the value, so that most of the mistakes are cumulatively less than £5,000, if you see what I mean. That's where the vast majority of mistakes are.

Q33 Mrs McGuire: That's what I wanted to get to. Is that £5,000 over a number of years, or over a couple of months?

Phil Gibby: It would be over quite a long time.

  Sir Leigh Lewis: When you get to it, yes.

Q34 Mrs McGuire: The question I'm trying to elicit an answer to is whether or not, in your opinion, in your experience of the system, there should be a de minimis in terms of error. We talk in billions, and it's a horrific figure we're going with, don't get me wrong on that. However, when you look at the individual cases, it's whether or not it is a cost-effective exercise, to start to try—where it's error; I'm not talking about fraud. I want to segregate fraud out from this discussion.

Sir Leigh Lewis: I think there's a lot there. First of all, just to say that my understanding is the same as the NAO's. When an error comes to light, the figure that's attached to it is the whole of the error since that claim was in error. It can go back, in some cases, for many years.

Secondly, we do try—I don't want to pretend it's perfect—for example, to assess the proportion of error and the size of that error for individual benefits. For example, pension credit is one example I know. We know that around two­thirds of pension credit error overpayment cases contain an error of less than £2 per week. We know that the equivalent for underpayments is 75%; the error is less than £2 per week. That doesn't mean it doesn't matter, but it puts it in relative perspective.

Q35 Chair: The quantum on that is that it's less than £2 a week in 75% of cases?

Sir Leigh Lewis: In terms of overpayment cases, it's £2 in 65%, and in terms of underpayment, it's 75%.

Q36 Chair: But the absolute figure is what?

Sir Leigh Lewis: The absolute figure I haven't got in my head, but it will certainly be in the Report somewhere. If I can't instantly find it, we'll give you a note.

Q37 Chair: Right. The £2 is minute, but it will end up being £70 million or something, probably. That's what I'm trying to get at.

Phil Gibby: The total of estimated overpayments on state pension credit in 2009-10 was £150 million; that's in figure 3.

Q38 Chair: So 65% is less than £2, so it's probably about £100­odd million. £120 million, I'd say. Something like that.

Sir Leigh Lewis: Just to finish the answer to Mrs McGuire's question, we do have, in terms of debt recovery, a de minimis figure. Not when it comes to fraud, as you say. That's simply good use of public money, that we don't spend more to try and recover a very small debt than we can get if we recover it.

Q39 Mrs McGuire: You do have a de minimis figure for error, then? You said that you don't have a de minimis recovery for fraud? I may have picked that up wrongly, sorry.

Sir Leigh Lewis: Debt recovery, where someone owes us a debt, we do have a de minimis figure, below which we do not normally seek to recover.

Q40 Mrs McGuire: And what's that?

Sir Leigh Lewis: That's agreed with the Treasury. I haven't got it in my head. John, have you?

Ruth Owen: Sixty-five pounds.

Sir Leigh Lewis: Sixty-five pounds. Ruth Owen, from Jobcentre Plus. Sixty-five pounds.

Mrs McGuire: So £65, if it's a pension credit of £2, that's the equivalent of what? Thirty weeks?

Chair: Thirty weeks.

Q41 Mrs McGuire: Is your £2 a weekly figure for the overpayment of pension credit, or is that a monthly figure?

Sir Leigh Lewis: No, that's—

Q42 Mrs McGuire: I'm trying to get a sense of whether or not the big machine is investing a lot of money to pull back a minimum amount, albeit that cumulatively it looks horrific.

Sir Leigh Lewis: The assurance I give you is that in seeking to recover debt—say, a debt owed by someone who has been overpaid pension credit—we would not normally seek to expend a great deal of money trying to recover a small amount. It is not good value for the taxpayer. We do, with both error and fraud, concentrate on the larger-value cases, inevitably and rightly.

Q43 Chair: Just to say on that, of course, in today's judgment, you won't be able to recover it anyway, if we're all right about what we heard on the radio.

Sir Leigh Lewis: That's absolutely right. The bulk of our recovery comes from customer error and fraud, rather than official error.

Chair: Okay.

Q44 Matthew Hancock: I'm very grateful—I think we're all very grateful—for your openness at the start in accepting how difficult this is, and perhaps your demob willingness to be very clear about that. I wanted, following on from that, to ask why, if you've been improving, as you were saying, in terms of reducing this error, it went up in the last year. If you look at the charts in figure 2, overpayments rose between 2008-09 and 2009-10, both in billions and as a percentage.

Sir Leigh Lewis: First of all, on the figures, when I said "improving", I think what I was actually saying was that I think our effort and our commitment have not wavered in all of that time. The figures have stayed pretty flat.

Q45 Matthew Hancock: Hold on. This is an increase from £800 million to £1.1 billion.

Sir Leigh Lewis: Yes.

Q46 Matthew Hancock: That's not "pretty flat".

Sir Leigh Lewis: That simply reflects a rising benefit bill. It's a constant proportion of a rising benefit bill.

Q47 Matthew Hancock: It's a higher proportion; it's gone from 0.6% to 0.7%.

Sir Leigh Lewis: The point I was making before is that it's fluctuated around 0.6%, 0.7%, 0.8% in recent years. The proportion of benefit being overpaid through administrative error has stayed pretty flat. The cash sum has gone up, because the amount of the total benefit bill has been going up year by year.

Q48 Matthew Hancock: Okay. You say that the nominal side is down to the increase in benefits paid, and you don't think the percentage is a significant rise?

Sir Leigh Lewis: I certainly do think it's significant; I'm not for one moment saying that overpaying 0.7% doesn't matter in any way. What I'm saying is that the changes in the proportion, over recent years—and I think that's the message of the NAO Report as well—are not significant, particularly when you include sampling error. The only conclusion you can sensibly draw is that it's remained flat, pretty well, as a proportion of benefits spend.

Phil Gibby: If I may just chip in to clarify, the estimates are based on a sampling of cases.

Q49 Matthew Hancock: Yes. I was just going to come on to that. The estimates are based on a sample from October 2008 to September 2009, and it says in footnote one in figure 3: "Complete data for 2009-10 will not be available until December 2010." Do you know when in December 2010? It might be useful to have a look at them before completing our Report.

Sir Leigh Lewis: I regret to say that it's not going to be until probably January 2011. This is not me. It is the statisticians who produce these estimates in our Department, and they don't release them until they think they're absolutely as accurate as they can make them.

Q50 Matthew Hancock: Okay. Can I just push you on a couple of factual points related to this? For the 0.7% figure that we were talking about in figure 2, where are the areas where the proportion of overpayment is lowest? If I look down this list of benefits, in figure 3, breaking them down, other than retirement pension and carer's allowance, which in total come to about £32 billion, all of those proportions are higher than 0.7%. I can't understand how, given that in big benefits, like housing benefit, incapacity benefit and income support, the proportion of error is significantly greater than 0.7%. I can't mathematically work out how that average can get to 0.7%.

Sir Leigh Lewis: The simple answer to that is retirement pension.

Q51 Matthew Hancock: Because it's so big.

Sir Leigh Lewis: Because it's so big, and the figure is 0.1%. Therefore that's the thing that brings your average down to 0.7%.

Q52 Matthew Hancock: Okay. Thank you. My final question is on figure 4. I just want to get some sense of these really big overpayments. If you're on income support, then an overpayment of over £5,000 is enormous to have to deal with as an impact on the family. An overpayment of over £10,000 is extraordinary, and the idea that you can end up in a situation where you've got an overpayment of £25,000 is very difficult to work out.

How can the Department get itself into a position where it has overpaid somebody by £25,000, when you're dealing with income support, which is about—you'll have to help me—£50­something a week?

Ruth Owen: Sixty pounds.

Sir Leigh Lewis: Can I give you an example of how this can happen—it might help—and also an example of what an overpayment is, if that doesn't sound almost too theological a question to ask in return.

The way this whole system works—and the NAO may want to add to that—is that because we can't go and examine every one of millions of cases every year, we take a sample of cases. We go through those, and the NAO go through them, in great detail to see if there is an error in a case. You may look at a case that has not been picked out ever in that sample process, or not for decades. If you then find an error in that case, in order to generate a value figure for it, you take that weekly or monthly error for the whole number of weeks or months that that has been in payment.

Take, for example, someone who's been long­term unemployed for a very long time. You might find, when you go back into that case, when it's examined, that it's in error, because there's no record that a Jobseeker's Agreement was ever completed. That's one of the conditions of entitlement, that there must be a Jobseeker's Agreement. If that person has been claiming benefit for, say, 10 years, you will then simply calculate that as 10 years' worth of benefit, because there is no evidence that there was ever a Jobseeker's Agreement. One of the ironies is that there may well have been a Jobseeker's Agreement. The failure may simply be that we never recorded it properly and it can't be found.

Q53 Matthew Hancock: On an overpayment like that, which is very significant to the individual concerned, how will moving to a universal credit help?

Sir Leigh Lewis: In that particular instance, that probably would not impact. Where a universal credit will undoubtedly help is that, going right back to the first answer that I gave the Committee, it will radically reshape the benefits system. All of those huge complexities, which not only our customers but our staff are wrestling with day in, day out, will become far simpler. There will be far fewer benefits. There will be far fewer premiums. There will be far fewer combinations. There will be far fewer sets of circumstance. The system will be a radically simpler one, and therefore the capacity of everyone to make mistakes ought to be dramatically reduced.

Q54 Mrs McGuire: We'll take it from a universal credit that there will be add­ons to the universal credit?

Sir Leigh Lewis: Yes, there will.

Q55 Mrs McGuire: So there is still, within the system, capacity for—

Chair: Complexity. Absolutely.

Mrs McGuire: Maybe we can come on to this later, about why the errors take place. The errors strike me as mainly personal; people not putting in the correct information. There's still the capacity for human error in the new system, because you can still have those add­ons. You might reduce it from 30 benefits down to 10 benefits, but unless you get the first page right, which is inputting the information, you're still going to have error in the system.

The universal benefit approach is a significant change, and I don't have any problem with that, but there's a danger of over­egging the notion that it will almost drive error out of the system, because I don't think it will.

Sir Leigh Lewis: No, and if I've given that impression, then I've given it wrongly, because you will never eradicate human error from the system. There will always be some levels of complexity, and that is absolutely true.

The Government's recent document on the universal credit, published on 11 November, estimates that the overall universal credit and related measures will reduce losses from fraud, error and overpayments by more than £1 billion a year in the long term. That includes, by the way, tax credits, which are not the subject of the hearing this afternoon, but will be part of the universal credit.

Q56 Matthew Hancock: Can I just ask one final question on this? The overpayment that you describe—where there's a lost agreement, therefore leading to a calculated overpayment of £25,000—isn't really an overpayment as we would understand it. It is a lost document.

How do you separate, in these figures, overpayments where money has been paid—where, in those circumstances, whatever the bureaucracy did, the money should not have been paid—from where the money ought to have been paid but the processes weren't followed? In the bureaucratic world, that's called an overpayment, but there's a difference. Do you see what I mean? There's a difference, a key difference.

Sir Leigh Lewis: Yes, I do.

Q57 Chair: I was going to ask the question in a different way. You've described what I would call a technical overpayment, rather than an administrative error. Have you got a proportion? Do you know how many are technical? What proportion of the £1.1 billion is technical? Do you know that?

Sir Leigh Lewis: I've got a figure that I do know, which is that we think in 2009­2010 one of the top 10 causes of official error was a labour market weakness in the claim, when it was found, and that typically would be no Jobseeker's Agreement in the case file. That was £33 million of error attributed to that specific cause in that year.

Matthew Hancock: That's not really an error, is it?

Chair: It's a technical error.

Q58 Matthew Hancock: The payment of that £33 million was appropriate. The error was that the documentation wasn't signed off.

Sir Leigh Lewis: You take me into an area that has occasionally been one of contention between ourselves and the National Audit Office, in fairness, because I think that we would ourselves argue that there are indeed errors and errors. There are some errors that are essentially technical and where it is not the case that you should have paid somebody £50 and you just put the wrong digit in and paid them £500. The National Audit Office—I am not complaining; it has been the system for many years, though I wish it were otherwise—takes a pretty austere view of this.

Chair: Just tell me, out of £1.1 billion, how much of this is technical? You've given us one example.

Q59 Matthew Hancock: How much should have been paid?

Sir Leigh Lewis: I couldn't give you that offhand.

Q60 Matthew Hancock: Amyas is going to tell me off, because I've been helping a Permanent Secretary against the NAO, which is not my job.

Sir Leigh Lewis: Why weren't you here earlier?

Q61 Ian Swales: Just to help the National Audit Office, the fact is, you don't really know, do you? That's the problem. You can say, "This document was missing, and therefore it's a technical error," and, as Mr Hancock says, therefore it's okay. But you don't know it's okay, because the document's missing.

Amyas Morse: Thank you. That's right.

Sir Leigh Lewis: Exactly. I think this just shows that there is some theology in what we count as error, and there always will be, if you see what I mean. There can be another variety of error that will go down as an error, where you pay someone a contributory benefit to which they weren't entitled, for example. Therefore that is an error, but actually, the likelihood is that they would have received the same amount of money on an income basis, so there are lots of technicalities in this area. It is not simple.

Q62 Dr Creasy: Sorry, I accept your premise that there might be a theological perspective to this, but there were a lot of books in the Bible, and I notice that you've got 607 different error categories. The idea that you don't have a category for these types of errors seems to me a little alien. You say that 52% of these were not used. What were the others, and why have you not been able to differentiate between these different sorts of errors? Obviously that's fundamental to the kind of issue you have to address.

Sir Leigh Lewis: Just to get a pause for breath, I might just ask Ruth Owen to take on that question, because we've been cutting down that number of errors precisely because it was too many to be useful and meaningful.

Ruth Owen: The description of the different types of error goes back to Sir Leigh's earlier statement about the range of combinations of things that could go wrong on a case, or a combination of benefits being paid together, and trying to categorise to find the root cause. What we've done now is reduce that down to about 60 different categories, of which we now focus all of our effort and error reduction activity on the top 10, to combine them together.

Q63 Dr Creasy: But the things Matthew's talking about are process errors as opposed to claim errors. Do you have categories for process errors? Would you be able, if you came back to us in 18 months' time, to say, "We've looked into our errors more, and some of these are about lost paperwork or contributory principles. Others are about inappropriate information"?

Ruth Owen: I have two sets of information. I have information that the NAO interprets, which is the sampling cases, but my understanding from Jobcentre Plus is that technical errors are about 10% of error that they found in Jobseeker's Allowance.

Q64 Dr Creasy: Do you have an error code for process problems?

Ruth Owen: Yes, we do. That's called "deemed error".

Q65 Mrs McGuire: "Deemed error"?

Ruth Owen: "Deemed error". That's where we have assumed an error because it couldn't be proved one way or another. Internally, within Jobcentre Plus, I have my own internal checking regime. We make 600,000 checks of benefit cases every year. I have data down from regional, local, site level, team level, to individual level, for all those cases that were checked, how many were paid correctly, so the person got the correct amount of benefit, and how many were procedurally accurate. I've got all that information, which is management information, and which I use.

Q66 Dr Creasy: Why haven't we got that information?

Ruth Owen: The NAO uses the information, which is the sample, and they have deemed error and non­deemed error.

Q67 Chair: I think we haven't got the information. What the NAO Report says is that you do do those checks, but you don't go behind that to understand why you have an error. I can't remember where it is in the NAO Report—

Sir Leigh Lewis: No. No.

Q68 Chair: —perhaps I can get help on that. In fact, your own checking system isn't used to best effect to understand where the error comes from, so is it technical or is it real, and for you then to take the action.

Sir Leigh Lewis: I plead, if not completely guilty, at least semi­guilty to that, although I don't think you can have a plea of "semi­guilty".

Amyas Morse: "Not proven", I think you'll find.

Sir Leigh Lewis: Sorry?

Amyas Morse: "Not proven".

Sir Leigh Lewis: "Not proven". The Comptroller and Auditor General's helped me out.

Mrs McGuire: Good Scots verdict, not proven.

Sir Leigh Lewis: What is clear is that we collect masses and masses of data. Ruth has given you some indication, but we collect data in all sorts of ways. What is also clear is that in the past, we've not, if you like, tortured that data. We've not really got the last ounce of what it's got to tell us about where those errors are happening, why they're happening, etc.

I might almost look to John Oliver to say something. That is precisely why we have established the Fraud and Error Council. John has now the most senior directors in the Department, both operational and policy, around his table in a group that reports directly to me. The whole purpose of that, or one of its purposes—and John, would you like to say something about this?—is precisely, are we extracting all the relevant information from the data we've got?

Q69 Chair: When you come in, you can tell me when you were set up, and when you went from this absurd 607 different error categories down to your 60.

John Oliver: I've been in the role now for a year.

Q70 Chair: Then it should be reflected in the Report, if you were in for a year.

John Oliver: I think there was some acknowledgement of what we are doing in the future in the Report. The review of where we are, our new strategies and our new programmes are, I think, alluded to in the Report itself.

If I go to the error code debate, and then we can move forward to that, you won't be surprised to know that that's a debate that's very much alive in the Department full stop. From operational staff right through into the corporate centre, this debate is alive. When is an error not an error? When is it deemed? Is this fair? Should it be included? What we have done is then reviewed all our error codes. Just before I came into role, we came up with the new figure of 60 error codes. However, we haven't stopped there, so we are now looking at reviewing them all again.

At this point, from our operational staff's point of view, it feels unfair that this deemed error doesn't reflect real error in the system, but is a procedural error—"Why does that reflect on me?" We're three-quarters of the way through a review, which we will be talking to our NAO colleagues about, about where we think we could redefine some of these error categories and some of these codes still further. What we want to do is take this debate still further in becoming a little bit more sophisticated about describing what an error is, what an official error is, within the system.

The problem is because we've got so many error codes, this is not just one error or one error code. We've been talking about hundreds of errors, because often there's no such thing as just one reason why a case may fall into error. You do have to reach a subjective judgment at times about what the basis of evidence is before you. That's why we had this whole system in the first place, the tick-box thing. What we're doing now is moving that on. Our review will begin debating, in the next few weeks and months, with the National Audit Office and within the Department, what we believe is a fairer and more accurate picture of what we deem to be error within the system. This will evolve.

Q71 Dr Creasy: You got 5,000 calls per month from staff seeking advice and clarification on how to handle specific claims.

John Oliver: Yes.

Q72 Dr Creasy: Last year you started to analyse the nature of those calls. What's the analysis going to tell you?

John Oliver: I'll hand over to Ruth, because that's a Jobcentre Plus­specific issue on what they are learning from their calls.

Q73 Dr Creasy: Presumably you're learning from Ruth? Presumably you've got Ruth's data in your Fraud and Error Council.

Ruth Owen: Yes.

John Oliver: A lot—we have lots of data.

Q74 Dr Creasy: So what are your staff telling you is the problem? If administrative error is as much the problem as claim error, what are they telling you they're struggling with?

Ruth Owen: Those calls to that advice line cover a real range of issues. "I'm sitting on the front line, I don't know how to process this claim." It can be, "I don't know how to use the IT system." It could be, "I don't understand the process I've got to follow."

We were specifically following up on the NAO recommendation about using that information. We've looked at what those calls are telling us in relation to guidance in particular, which is quite a small proportion of those calls. A key message that we've taken out—and I have followed up afterwards to talk to staff about—is about the clarity of the guidance that we have.

We have all our guidance on our systems, and you can type in and look at: "What do I have to do with this particular case, because it's complicated, it might have a particular type of mortgage, or something, and I want to know exactly what I've got to do." The staff are saying, "All the information is there, again, complex though it is. It's not very easy to navigate." The big headline I took out of the research into the calls for the feedback from the advice line was that navigation of our guidance isn't good enough to help our staff get the right answer.

Q75 Dr Creasy: How do you think that's going to be affected by bringing in another system for payment of benefits?

Ruth Owen: It removes quite a lot of the other guidance that we wouldn't need anymore.

Q76 Dr Creasy: But it means there'll be new guidance, with more additional components to it.

Sir Leigh Lewis: Yes.

Q77 Dr Creasy: What's your capacity as an organisation to learn and grow? What this data seems to be telling you is that you've got quite a substantial problem with your staff not understanding the systems that you're operating. What's your sense of your ability to be flexible and adapt to it?

Sir Leigh Lewis: I do think we have a problem with our staff not understanding the systems. That takes me right back to the complexity of what we're asking them to understand, and, in fairness to ourselves, what Governments and Parliament have asked them to understand, that huge wave of complexity. Without, again, putting forward universal credit as the answer to every issue, that will radically simplify; but, look, as you've said to me already, that's a long way away.

Q78 Dr Creasy: It's also got its own differentials within it, as well. It's not a completely singular service.

Sir Leigh Lewis: No. I absolutely accept that.

Q79 Chair: Local rent allowances, to raise something contentious.

Sir Leigh Lewis: There will always be complexities left in the system. On what we are doing in the here and now, Ruth's message is clear; there is an awful lot of guidance out there, but you have to go back to an individual member of staff seeing a customer, or talking to a customer on the telephone. These are busy and pressured jobs. What that member of staff wants to be able to do, if they don't understand what they should be doing, or they don't understand the rules, is to be able to find the piece of guidance quickly that tells them what they should do in these circumstances.

That guidance is all there—it's all there somewhere—but it's not very easy to find at the moment, though we've made improvements. I'd like the opportunity to say this, in response to the NAO Report: I've commissioned—perhaps it is the last thing that I will commission—a review into our guidance, as a result of this Report, and asked it to report its conclusions to my successor—on how we can make our guidance clearer, easier to access and more user­friendly.

Q80 Stephen Barclay: Building on the same point, you've just said you've commissioned a review, but I assume you had a review three years ago. In the 2007 Getting Welfare Right document, you talk about reinvigorating performance: "effective organisations learn from their own performance. They also monitor staff performance proactively, and in order to align ourselves more closely with this model, the Department will go further than it has before to improve its performance."

You were talking about that three years ago, and yet this Report, as Stella was touching on, is saying that until you saw this Report, you were getting 5,000 queries a month from staff, saying that they did not understand something. Yet you weren't using that information whatsoever to design the IT platforms which they themselves are using.

Sir Leigh Lewis: We can talk about the IT, because I wouldn't go all the way. If what you're saying to me is that some of this you could have done earlier, harder, faster, and you could have put in even more effort, even more resources, even more commitment—yes. When you look back with hindsight, there's always something extra that you could have done.

Let me just take a couple of quotes from the report, in fairness to ourselves. The Report says: "The Department published its strategy Getting Welfare Right: Tackling Error in the Benefit System in 2007 and there has been a regular and constant management oversight of progress." In a different quote, "The Department and its senior officials are clearly committed to reducing the cost of administrative error". This is something that has been on my personal priority list for the last five years.

Q81 Chair: If I can just come back at you with one or two quotes, one from the 2007 document, one from the 2010. Stephen's probably got other ones. In the 2007 document, you said you would "pursue implementation of a system of formal accreditation of skills".

Sir Leigh Lewis: Yes.

Chair: In the 2010 one, it says DWP will run a system of accreditation of staff.

Sir Leigh Lewis: No, but actually, Chair, some real progress has been made since those two periods. We have launched a programme of accreditation for our decision­makers, in the Pension, Disability and Carers Service. Those are the people who have to take some of the most difficult decisions on those benefits, like attendance allowance, DLA, etc. We now have some 300 of those decision­makers fully accredited, under a programme that we have worked up and delivered with the University of Chester, and that is an accredited programme that has recognised qualifications at the end of it. Now we're moving further, but we have taken some big steps on accreditation.

Q82 Stephen Barclay: It puzzles me; it just feels a bit like "Groundhog Day". We're getting the same announcements as per several years ago. We've got 5,000 staff phoning up saying they've got a problem. We've got the Department having an IT system that's not designed from the point of the user. Again, in 2007 you were saying, to quote, "we will strive to design new computer systems and business processes from the perspective of those who have to use them on a daily basis." And yet, on 29 November 2010 you are saying "By designing new systems from the perspective of the user, we will help to make sure that they prevent staff from making errors."

In your evidence earlier, you were saying to us that it's important to stop administrative errors from getting into the system, and I think we would all agree it's a lot easier to stop the problem than to solve it. But you've got this wealth of information from your own staff phoning up. The NAO is going in and finding issues such as that your staff think the training's inadequate, the guidance doesn't help them when they need it, support for staff is not readily available and they're working to conflicting targets. I just can't see what has changed between now and three years ago, when, with respect, you were saying pretty much the same things.

Sir Leigh Lewis: Let me give you some very specific examples on exactly what's changed in relation to the IT. I can go with you on training and guidance, and I can tell you what's changed there, because this has been a constant journey on which we have been taking action after action.

Two fundamental exercises, which we held around 2007 to explore how to tackle the top errors in income support and JSA, led specifically to two major IT enhancements, which we introduced at the end of 2007. These were to address major causes of error where the customer, the person claiming the benefit, is no longer a lone parent, or is no longer sick and their circumstances have changed. The enhancements now automatically issue a letter to the individual, telling them, in the case where a child is passing a certain age point, that their benefit is due to finish when the youngest child reaches that age. Having done that, the system then suspends the benefit automatically at that point, whereas previously it required human intervention to do that. Those are both examples, and I have others here which I could quote to you, of where we have been using our IT very rigorously to design error out of the system.

Q83 Stephen Barclay: So in terms of your major IT projects, how many of those are currently late or over budget?

Sir Leigh Lewis: Very few, actually. The Department has a very good track record of delivering its IT programmes successfully. The most recent example of that was the Employment and Support Allowance, which was delivered to time and to budget. The great majority of the Department's IT programmes have come in as planned and on budget.

Q84 Stephen Barclay: The 2008 Report that said that 19 IT projects at the Department of Work and Pensions were over budget and late was incorrect, was it?

Sir Leigh Lewis: Well, I haven't got the Report that you're quoting from there.

Q85 Stephen Barclay: This was from the Work and Pensions Select Committee, which identified, "Out of the 19 identified by the National Audit Office as being the most significant, nine are expected to be over­budget and ten are late." This is an article in Computer Weekly on 19 November 2008.

Sir Leigh Lewis: With respect, I'd like to go away and check the Computer Weekly figures.

Q86 Stephen Barclay: Sure.

Sir Leigh Lewis: Not everything that Computer Weekly has ever published about Government IT systems has always been wholly accurate.

Q87 Stephen Barclay: I accept that, but you're saying that the track record is very good. They're saying that all the 19 big ones are over budget and late. It just seems there's a big gap between the two.

Sir Leigh Lewis: Why don't I drop the Committee a note on that one? It's quite a way from today, but we do have a very good record.

Q88 Jackie Doyle-Price: I'd like to address Ruth, if I may, on some of the findings regarding your service. To be fair, I think you've moved quite a long way fairly quickly. You looked at the targeted checks, found they weren't getting very much return, so by 2010, you'd actually moved on and refocused your efforts to find that most overpayments arise when processing changes in customer circumstances rather than new claims.

If you take a step back, that seems pretty obvious. It does raise the question as to whether we're obsessing too much about data for a very small number of cases, when you need to challenge your overall systems more. I'd just like to invite your observations on how you got from those random checks to where you are now, and perhaps how you're going to address what you have found in future.

Ruth Owen: You're absolutely right. I think we first focused on new claims because of the data: to capture a new claim correctly you have to capture, particularly for a means­tested benefit, quite a large amount of information. Our assumption was that a lot of error could fall into the system at that stage, and our strategy was to keep the gateway to benefits quite tight, by being really rigorous about not letting error through at that stage.

You're absolutely right, however. The more you dig into it, the more you realise that as benefits progress, either things don't get corrected, or things don't get uprated at the right amount, or customer information doesn't flow through to us either continuously or accurately. As people are on benefits for a certain amount of time—people are on means­tested benefits, in particular, for several years—they can drift into inaccuracy either through official error or through customer error. We have definitely changed the way we target our checking regime to make sure we're focused on the high areas of risk.

What I do now, every quarter, is look again at what our root cause analysis is telling us, to say: "Right, now what do we know?" Something drifts to the top of our top 10, we knock it off by doing an IT fix that Sir Leigh's described, and something else rises to the top. You then need to create a new checking regime, to say, "What is now the biggest cause of error that we should be focusing on?" Every quarter, my team now write out to the whole business to say, "Our latest top five errors that you really ought to be focused on are these, and this is where I want you to do the vast majority of the checking."

Q89 Jackie Doyle-Price: What proportion of claims that you have to process are changes in circumstance as opposed to new claims?

Ruth Owen: Off the top of my head, it's a ratio of about one to four, something like that. I can just think of the weekly figures that I see. Roughly, on a weekly basis, we do four changes for every new claim—something like that.

Q90 Jackie Doyle-Price: So you've then got to segment those, and obviously you're going to be more effective at tackling error in the system if you've focused effort there.

Ruth Owen: The load, basically. We either look at the flow—how many things are coming into the system—or we focus on the load. Primarily what we're saying now is focus on the load, both concerning change of circumstances but also on the "spend to save" investments that we're talking about.

We can do "case cleansing", as we call it—diving into the load of a particular benefit and making sure we check every single case with a particular set of risk categories. We will look at them, literally phone the customer up and say, "Are these still your circumstances?" and make sure that the case is still correct. That's what a lot of that investment is to do.

Q91 Ian Swales: I'd like to return to something that Mr Barclay mentioned a few minutes ago, which was the items on page 23. At the bottom, there are a number of items about why errors occur. Clearly, we know that people are at the back of this. I must say I was very impressed with the people I met at my local Jobcentre Plus last week. Clearly those good people need to be enabled, and I particularly want to draw attention to the bottom two paragraphs, which are quite damaging, when you read them. "A belief that training is inadequate—learning and development for new staff was criticised for poor delivery; training facilitators lacked technical expertise; training packages were described as being out of date, incomplete and did not mirror the real work."

That's quite a damaging assessment of training. That's one area. The other area is the issue about the guidance, but maybe I'll talk about training first and just ask something about the guidance.

Sir Leigh Lewis: I'm going to say something about the training. We invest a huge amount in training, an absolutely vast amount in training. It's not a lack of investment in training. Training is one of these things where probably you never get it entirely and totally right. I will give you some sense of the scale of what we do, and then perhaps I'll ask Ruth to say a little bit more about the training that's provided on benefit accuracy and benefit delivery.

This year, Jobcentre Plus will deliver about half a million training days. That's about an average of five for every member of staff. Our other major business, the Pension Disability and Carers Service, will deliver over 100,000 training days. The way our training system works is that, for people newly coming into the organisation or moving from one specialism to another, there are accredited periods of initial training. They're normally then followed up by a period of consolidation, where people work with a more experienced member of staff, a buddy, and so on, and then there's a checking regime.

Our staff are not ecstatic about training, and that is the truth of it—the Report picks up what I pick up, what Ruth picks up, and John, as we go round. One of the reasons is that, like many other organisations, we've moved—not completely in any sense, but to quite a degree—away from the classic "chalk and talk" to e­learning. There is still a residual fondness for the chalk and talk in the organisation—taking people away and putting them in a classroom for a week, for two weeks. We still do an awful lot of that, but we do much more e­learning, and that isn't always the most popular.

Q92 Ian Swales: It's an impressive answer in terms of quantity, and ethos, if you like, but the comments are that training packages are incomplete, out of date and did not mirror the real work. That's inexcusable. It's even less excusable in an e­learning environment, where hopefully you can keep up to date in terms of the real jobs that people are having to do.

Sir Leigh Lewis: This is one—and then I will hand over to Ruth—where in one sense perceptions are everything. If that's what our staff are telling the NAO—it's a fairly small sample, in fairness, but nevertheless—you take that seriously. The truth is, that doesn't necessarily prove that all those things are true, or that those things aren't covered, or the training is out of date, etc. It is a perception. It is a perception at times when people are terribly busy and working terribly hard, and the Report gives us the credit, actually, for the way we cope with huge rises in volumes during the recession.

Q93 Chair: Before we come to Ruth, does the NAO want to come in on that? Is it your view that it's real or is it a perception?

Phil Gibby: It's very difficult; we do caveat it, because I do take the point that if you go and ask staff, to be honest, people will always complain about something, if you see what I mean. You need to temper it with a bit of caution. Why we've got it, and why we think it's important, is that with a lot of the error codes and things we're talking about, the analysis is by type of error. What you need to do is get to the underlying reasons why the mistake has arisen. If the wrong number has been put in on the computer system, is it because they felt under pressure and they didn't know what they should be doing? It's about trying to unpack and get to the bottom of it.

Q94 Stephen Barclay: How does this Department compare with other Whitehall Departments on these sorts of issues—training, staff morale? How do you benchmark them? That's another way of looking at it.

Phil Gibby: That is difficult to measure, to be honest. The Capability Review covers some elements of those, and I think the Department comes out well.

Stephen Barclay: There's a satisfaction survey, for example, that covers all Whitehall Departments.

Sir Leigh Lewis: I can give two pieces of information on that. Capability Reviews have been the most systematic process, whereby the Department has been assessed externally against a whole range of indicators. The Department has emerged very well indeed from those in relation to Whitehall Departments generally. There have been two rounds of reviews; they are all published on the Cabinet Office website. The Department has emerged very well. That measures the Department against a whole range of performance indicators.

Q95 Mr Bacon: From memory, it was 10 indicators across 17 Departments. Overall—I remember the figures—two­thirds of Departments were less than well­placed. You're saying the DWP was better than that. What proportion of your gradings were well­placed or better?

Sir Leigh Lewis: I just haven't got that in my head.

Q96 Mr Bacon: Could you send us a note?

Sir Leigh Lewis: Of course I can, and I'd actually be quite pleased to, because you'll see we emerged in the top two or three Whitehall Departments from those Capability Reviews. On the question about our engagement survey, this is absolutely a half­full or half­empty glass. Our engagement figures—and we participate in the annual Civil Service survey—are absolutely not figures to die for. The absolute levels are much lower than I would want, and that's, in part, a feature of very large organisations, which tend to have lower engagement scores than smaller organisations. Up until this year, we have been improving for four successive years. If you take almost every score, it's been going up, but from a low base.

Q97 Chair: Just to bring us back, Ruth, on the training.

Ruth Owen: Very briefly, then, I do recognise what the NAO has described here in anecdotal feedback, but I really don't think it's as dire as it's portrayed. If you think that we recruited 16,000 people over the last 18 months, during the recession, I went out every week of that year, talking to people. I heard lots of complaints, mostly about recruitment processes. I did not hear a single complaint from staff about training. We evaluate our training, and training satisfaction from our staff. Some 81% of staff said they're satisfied with the training that they get.

There is a genuine point about methodology of training, about people wanting classroom­based training. I suspect it's people like me; when I joined the organisation, I went for 26 weeks of endless training without going near a real case. I suspect a lot of people have a fondness for that method of training. What we do now is modular training. We get a couple of weeks on this particular benefit, then you go and sit with somebody and you consolidate on that with real cases, and then we grow you bit by bit. I don't think it's fair to say either that the training was out of date or that it doesn't relate to the real world.

Amyas Morse: Can I just be sure that we don't pass this point by? We very carefully caveat this. We're not presenting it as being any stronger than it is. The question we asked was, "Why do you think official error occurs?" We didn't ask "Are you happy with your training?" That was not the question, if I may just make the point clear to you. The question was, "Why do you think official error occurs?" This is what they volunteered. We're suggesting it as indicative for further inquiry, which we very definitely think you need to make, because the figures aren't getting any better.

We're not arguing with whatever your satisfaction stats are, but it's not incredible that the effectiveness of training may be a factor in official error levels. I guess you'd agree. We're simply suggesting areas of inquiry.

Sir Leigh Lewis: I accept that, as you know. That just seems eminent good sense, and it's an area we should look at.

Amyas Morse: Thank you.

Q98 Ian Swales: Could I just make the other point about the guidance? Clearly, this is a complex area; I recognise that. Are you investing enough in what you might call internal IT, so that people are able to access this guidance, maybe even through intelligent systems that enable them to get to the answers they need quickly? I hope it's not literally hauling big books off shelves all the time. So, can you give us some confidence? Just picking up the NAO point, the comment at the foot of page 23, specifically about why errors are occurring, refers to the guidance.

Sir Leigh Lewis: I think guidance is really important in this because you can have all the guidance you like but if people can't access it and understand it then it might as well not exist. Just to give you some assurance and comfort, there are still some great volumes on shelves for decision makers—the really, deeply complex stuff—but the great bulk of our guidance is now on our internal intranet, as we call it, our internal IT systems.

We've brought, for example, all the Jobcentre Plus guidance together into a single location on that site; before, it was spread all over the place. You can now get to the top 15 areas of guidance that people most need from a single tab on the home page. I could go on, but I'd bore you to death. So we've made a huge amount of progress in trying to make our guidance successful, but we still get feedback that our guidance is hard to find, hard to locate and not always written in the simplest terms, which is why I've said that I'd like a review of that.

Q99 Mr Bacon: May I come in at this point, because you did say earlier that the answer is radical re­engineering? I'd like to come on to that specifically in a minute, because that presents all kinds of questions of its own. But it speaks to the question of the idea that better guidance isn't going to be enough, because it begs the question of what the guidance is about and if the thing that you're giving guidance about is too complicated then plainly you're not going to get the answers that you want.

When he originally took over the Post Office, Allan Leighton went into work in a sorting branch. I remember hearing him interviewed on the radio afterwards and he said he wanted to do that to talk to people on the ground, because, he said, "Fundamental change is not something you do to people; it's something you do with people."

I wrote to you after a lady came to my surgery a few weeks ago who'd had 34 years' experience in the DWP, DSS, DHSS across all benefits, covering both a sink estate in North London and the more affluent areas of Norfolk and Suffolk. She'd got a lot of experience. You very kindly wrote back to me, and you dealt with the very specific issues that she raised in her meeting with me. But her letter more generally, which went into a lot more detail, is just replete with a sense of deep experience and frustration, which is the reason why I sent it to you.

When I was a journalist and when I used to represent the consulting industry, I used to write about turnaround in organisations. The most successful organisations, when they are changing fundamentally, most often do so most successfully by a higher level of staff engagement. Now, I'm not telling you anything that you don't know; I'm not telling you anything that expensive consultants haven't told you, but the question is how you actually do it. So can you speak to that a little bit? Then at some point before this hearing ends we ought also to come on to the whole question of the radical re­engineering you're talking about as well.

Sir Leigh Lewis: I remember the letter, because I remember writing the reply and thinking about it quite hard, having read your constituent's letter. So I absolutely remember that.

Two things: one, if you're entitled to be proud of anything, perhaps pride comes before a fall. One of the things that we have instituted to try and sort of get close to this—because there's always a terrible risk that those at the top of organisations are completely losing touch with what's going on at the coal face of the organisation—is probably the biggest back­to­the­floor programme of any in Government, if not actually in the private sector. We've asked every single member of what we call the Senior Civil Service—that's the top 300 people in the Department—to spend a week each year trying to do the job that our staff have to do in one of our customer­facing businesses.

Because you can't ask people to do it and not do it yourself, I've done that four times, in four successive weeks, including working for a week in York Jobcentre Plus, doing exactly that kind of job: signing people on as they came in every fortnight. Ruth and John will be able to tell you where they've been.

The really important thing is that almost everyone who goes out comes back with at least one frustration that we didn't know about and that the staff are fed up to the teeth with, and somehow in a large organisation no one has been listening. We use that learning to try and not just change those specifics but also find out generically what's going on.

One other thing which I do think talks to this directly is that we've instituted in the last few years a major programme—it's a jargon word, I know and I was never very fond of it, but I'm a bit fonder now—of Lean.

Q100 Mr Bacon: Lean?

Sir Leigh Lewis: Lean. Lean engineering.

Mr Bacon: Oh, Lean.

Sir Leigh Lewis: It came out of the motor industry originally.

Q101 Mr Bacon: Kawasaki and all that?

Sir Leigh Lewis: Yes, absolutely.

Mr Bacon: Kawasaki for people, so to speak.

Sir Leigh Lewis: Toyota was the great originator of this. Stripping away all the jargon, Lean is basically very simple. You do need to give your staff the tools, the techniques and the training, but you ask them to look at the processes you're using from soup to nuts—from the very beginning to the very end. You invite them to tell you how those processes can be done much more effectively and efficiently.

Again, we've absolutely gone for that in a major way. We've got 1,000 of our staff trained as Lean assessors and we have introduced lots and lots of changes. You just find there's a process with 400 steps in it to get from this point to an outcome, and, as you look at it, in the end you realise that 150 of those steps are either adding no value whatsoever or are positively counter-productive. So absolutely, it's really hard in a big organisation to have your finger on the pulse out there, but we don't half try.

Q102 Mrs McGuire: Can I go back to these 5,000 calls a month that you get? What do you do with them?

Ruth Owen: What do we do with them?

Q103 Mrs McGuire: Yes, I think earlier on you said that you had 5,000 calls a month. I think I'm right. What happens with those individual members of staff who have made those calls? What do you do? What's the process?

Ruth Owen: The immediate thing is we solve the problem that they've called up about. So they phone up because they've got a particular problem. The vast majority actually relate to the Social Fund.

Q104 Chair: Do you use the information?

Ruth Owen: Yes we do—albeit belatedly, I think it says in the report. We've only been collecting management information since earlier this year, so we start to get more feedback from building up that management information. So on the types of calls we're getting, I can tell you that the vast majority are about the Social Fund rather than about what we're talking about today. But of those that are relevant to today, the analysis that was done was that the guidance was something that people were specifically ringing up about.

Q105 Mrs McGuire: Do you analyse where the calls are coming from? Are they from groups, clusters of offices or individuals, or whatever? Collecting information is one thing and then using it to improve the quality of the service is another. What else do you do with it? You can tell me it's the Social Fund, fair enough.

Ruth Owen: Now, I don't think as far as I'm aware that we look at it in terms of where the calls are coming from. What we do is look at it in terms of who's got the solution, so we look at which parts of the Department cover that. Who's asking the question and therefore do we know what the problems are that are causing it?

Q106 Chair: I think the purpose of all this questioning is: are you using it?

Sir Leigh Lewis: Yes.

Q107 Chair: It's information that would be useful to understand what's going wrong, because these guys ring you because they can't cope.

Sir Leigh Lewis: Can I ask John to say something about this? One thing that we're doing now, in a way that we haven't done previously, is in John's group we're trying to do, benefit by benefit—and the first one we've taken is pension credit—something called a "diagnostic". That is another slightly jargony word, but essentially it means taking in, in a really structured way, all the information we've got about what is happening in relation to a particular benefit. Testing, analysing and then coming up on the basis of that information with a plan of action, saying, "Right, these things need to be tackled."

John Oliver: The reason why we call it a diagnostic is that what you find is you can have 101 different initiatives going on within each benefit—or within, indeed, the Department to look at problems within the benefit, whether it be error, fraud or whatever. What we are now doing is pulling together one work programme. The one work programme looks at all benefits and essentially is governed by the Fraud and Error Council, which I chair. The council is represented by directors across the Department, who can make decisions on where we are. Essentially, what we do and what we've done now is pull together all our investment across what we are doing to try to control, reduce, improve our performance on benefit fraud and error and then subject it to some challenge.

So for us, what we're looking at is saying, "So, what is the business case?" because for every piece of activity we should have a return, and I think we've been rightly criticised in the past for not using enough feedback information to say, "So, is that the best thing that we should be doing?" Now we have one programme for every benefit.

Q108 Chair: So you use this information?

John Oliver: Exactly. Not just this information because we've got lots of change activity; lots of investment that we have made across the piece here. There are lots of things that we are trying to improve on—whether that be IT systems, retraining or dedicated resource. All those then find their way on to the programme. We can then look to the root cause. So what we are interested in is saying, "If we stand back from this benefit, if we look at it from an independent point of view, are we addressing the fundamental problems?"

Now, the hotline is very specific detailed stuff. What I'm looking at from a strategic point of view is what the Department and our delivery agencies responses are to the big problems that we've got within each benefit, and are there then any gaps from within them. What we're then doing is subjecting that to regular challenge so that we can see if there are any gaps as we go through.

Q109 Mrs McGuire: What that might throw up is the strategic direction—

John Oliver: Indeed.

Q110 Mrs McGuire:—if you analyse the information. I suppose what we're trying to get at is what exactly do you do with this plethora of information that the DWP has, both internally and externally.

Can I just ask one follow­up question? On page 28 paragraph 3.11, the report indicates that two of the agencies, Jobcentre Plus and the Pension, Disability and Carers Service, have looked at different ways in which they can improve performance. I think it's quite interesting that under Jobcentre Plus it has indicated that there could be the potential to produce "perverse results with the more competent quality checker identifying more administrative errors than the less experienced," and identifying that "some staff may be reticent to identify errors". Then with the Pension, Disability and Carers Service, it says that their approach, which appears to operate independently, has produced more consistent results because of that independence.

What exchange of ideas has there been between Jobcentre Plus and the Pension, Disability and Carers Service if one model appears to be delivering a far more consistent product than the other?

John Oliver: Exactly. That's exactly what the council is about. It's about making sure that we share that best practice. For example, the recent strategy for fraud and error that we've just published is about benchmarking not just actually between PDCS and Jobcentre Plus but also the tax credits organisation and HMRC too.

Q111 Chair: So on that specific one, is Jobcentre Plus now doing its checking for people who are not in the Jobcentre Plus office now?

Ruth Owen: We are going to move to that model.

Q112 Chair: When?

Ruth Owen: As soon as we can.

Q113 Chair: When?

Ruth Owen: Within this operational year, I would move to that model. What I need to learn the lessons from is the PDCS, which has local checkers and independent checkers. At the moment I have local checkers and I have people who validate that separately.

Q114 Chair: So by April next year you will have independent checkers?

Ruth Owen: We will have some form of independent tier, yes.

Q115 Chair: Some form of?

Ruth Owen: Independent tier of checking.

Q116 Chair: What does that mean? Anne's point is a very good one. You're learning in one bit of your system that something works better. Why aren't you replicating it in another bit of your system?

Ruth Owen: What I'm saying is I want to learn from what the PDCS is doing and I've agreed to do that because I do think that having independence does address that perverse behaviour.

At the same time, there are also benefits to the process that we've been using, which is about local ownership and having that regular feedback loop. I was in a benefit centre this week talking to the checking team who talk to the team leaders on a daily basis about the feedback that they're getting on the ground. So there are pros and cons to each way; what I want to do is to try and achieve both within a resource. I don't know what I've got for next year yet.

Q117 Mrs McGuire: That doesn't detract from the fact that we have independent checking; you have consistency of performance. If I can remember correctly, the Pension, Disability and Carers Service used to have one of the highest satisfaction rates of any service. So the service has been delivered competently and effectively and yet they are still maintaining a more consistent result in terms of administrative error. I don't think the two are counter-productive, frankly.

Sir Leigh Lewis: I think you're on to a very good point, both you and the Chair. One of the things we've always struggled with in the Department—and, if I may say so, both of you will know this—is that we have separate agencies and they give you lots and lots of strengths, because they have an ethos, a belief, a commitment and a passion and they want to do their job well and they are doing different things.

But one of the downsides is that that is not invented here; that it can be, "Well, if we haven't thought of it ourselves, it can't be the right thing to do." That can spread more widely, even within agencies and within parts of the Department. What we're trying to do, very, very much indeed, is ensure that we really do share learning and we do share experience, and where something is working well in one part of the organisation and will transfer, we transfer it.

Q118 Mrs McGuire: Yes, but you lead together as a senior management team of directors. Frankly, with the greatest respect, to say that nobody's going to do it this way in this part of the organisation—I think it's taxing us just a wee bit to say that Jobcentre Plus or whatever other element of DWP is one organisation; you have said one organisation. Surely there should be a mechanism by which at a senior level they are encouraged to look at the good practice in other parts of the organisation?

Sir Leigh Lewis: Yes, absolutely. And come on, you know I didn't use the words that you just put back as if I had used those words.

We absolutely spend as a senior team a huge amount of our time learning from experience on this and in all sorts of other areas. But it is nevertheless the case—it's just part of organisational cultures—that if you have organisations, they tend to live within their own frameworks and boundaries, and I think all organisations that operate on a federal basis struggle to ensure that best practice is identified and spread.

Q119 Chair: I don't think we got from Ruth Owen an unequivocal, "I'm going to do this."

Ruth Owen: No, I think you did.

Q120 Chair: I didn't hear it quite.

Ruth Owen: I will set up an independent tier of checking in Jobcentre Plus by April. Is that clear?

Q121 Mr Bacon: April of which year?

Ruth Owen: 2011.

Sir Leigh Lewis: Is there something about trust here?

Q122 Matthew Hancock: It's just good to get it in black and white. I want to press you on the Fraud and Error Council, which you've spoken about a couple of times. If you could turn to page 10 in the report, paragraph 22. You talked a lot about how the Fraud and Error Council is going to ensure that you're going to reduce the cost of administrative error. Do you recognise the second, third and fourth sentences of paragraph 22, specifically that "The Department… lacks the information needed… has yet to develop a consistent method… and lacks costing data," and "accordingly… does not have sufficient evidence to demonstrate that existing activities to reduce administrative error have been value for money"?

Sir Leigh Lewis: Let me say a word or two. We inevitably, as you do, discussed this report at some length with the National Audit Office and, in doing that, with the Comptroller and Auditor General. I think that overstates it by quite a margin. I do not think that we are in that position. What I do absolutely accept is that in each of those areas there is more and better to do.

Q123 Matthew Hancock: Okay, so Mr Oliver, your council has been in operation for a year?

John Oliver: Yes.

Q124 Matthew Hancock: Since which month?

John Oliver: January.

Q125 Matthew Hancock: Since January. However, do you still feel that you lack the information that you need to target initiatives effectively?

John Oliver: Yes and no, and I'll tell you why I give that answer on both. Yes, I think that the easiest thing in this world is to be able to say, "I can look at the gap. I can look at official error in income support and this is what it costs you." I can then start drilling that down and saying, "This is what we think the problems are," be it mortgage costs, earnings entitlement, signing on, or whatever it's going to be, according to the benefit.

If you then drill down further and further, which is where we get into a debate both within the Department and indeed with the National Audit Office, then it gets more and more difficult to get more granular and granular about what it costs you and about what the response actually is. I think part of my job is to get the Department and get us better at being able to understand that far more definitively.

Where I think we can be criticised is at the moment I can say, "Yes, I understand how much we spend on improving one benefit." If you then said to me, "Let me go right down to the granular root cause," and say, "Yes, I really, really think I've cracked that," at Jobcentre Plus and PDCS level that's where I get into, yes. I think those delivery agencies will say, "Yes; yes, we do understand what the problems are." From a departmental point of view, from looking at it from my strategic hat, then I say no, there's more that we could do about becoming more sophisticated and understanding the return on that.

Q126 Matthew Hancock: That answer is perhaps not surprising, but if it were the case that you had the information that you needed, then wouldn't you therefore have managed to reduce administrative error more, or indeed reduce it as a proportion? It has gone up as a proportion by about 13%, from 0.6 to 0.7.

John Oliver: The important thing that I learned when I came into this role is that there is a time lag between investing and getting a return on your investment.

Q127 Matthew Hancock: Have you yet managed to develop a consistent method of assessing the cost effectiveness of your interventions?

John Oliver: Yes. I haven't got it so that I can produce to you now and say, "This is what it is", but I think that, yes, my analysts are coming up with a much more robust approach to being able to say "yes" to that question and being able to then put that before anyone—this Committee or the National Audit Office—and say, "This is our story."

Q128 Matthew Hancock: Obviously, these are the high-level conclusions.

John Oliver: Sure.

Q129 Matthew Hancock: In paragraph 16 on page 8, it says: "Jobcentre Plus reviewed nearly 541,000 transactions in 2008-09…but the information collected does not enable comparisons between different parts of the delivery chain." This is part of the reason that supports the conclusions that you don't have sufficient evidence to make comparisons of where the interventions are cost-effective. Has that changed since 2008-09?

John Oliver: Yes, I think our understanding has moved on since 2008-09 and I think that in our year of investment we do understand much more precisely what our return on investment is.

Q130 Matthew Hancock: You've said that about a broader point. The point raised in paragraph 16: does the information that you now collect enable you to compare between different parts of the delivery chain?

John Oliver: Yes. This is a specific Jobcentre Plus point and I know Ruth wants to make a point.

Ruth Owen: The answer is yes. We have three parts of our operation. We have contact centres, which collect information over the phone from customers. We have job centres, who are generally doing the face­to­face interventions with customers, and we have benefit centres. What that's trying to describe is: do we know which part of the business the error arose in?

The answer is now yes, we do do that. For every error that we look at, we try to find the root cause and we've introduced new checks into the contact centres and the job centres, because it's not all error manifesting in benefit centres. The benefit centres can look back and say, "Was that something that raised an error." As soon as that phone call was made, it's made its way all the way through the system, so we now know where the sources of error are, and I monitor that on a monthly basis.

Q131 Matthew Hancock: When did that happen?

Ruth Owen: Earlier this year. I couldn't put a month on it, sorry.

Q132 Matthew Hancock: Why in this report, which was agreed between you, presumably, Sir Leigh, and the NAO, does it say, "The information collected does not enable comparisons between different parts of the delivery chain"? I've just asked you, "Does the information allow these comparisons?" and you unequivocally said yes.

Sir Leigh Lewis: Let me take that and maybe the Comptroller and Auditor General might want to say something. Previous Comptrollers and Auditors General had a vision in which reports went on forever, if you see what I mean, through 47 different drafts, until everyone had signed up to the last comma and the last semi­colon. We are in a different world, and I do not complain about it, in which the Comptroller and Auditor General seeks views and then reaches a view. We exchanged quite robust comments about some aspects of this report.

Phil Gibb: If I could chip in. I think Sir Leigh's right; we did obviously have differences of interpretation in terms of conclusions on this. From our point of view, in particular, for example, on paragraph 16, it does relate back to the issue about having an independent checker in JCP. I take the point that you can get a division between the contact centres and the delivery centres, but actually you need that kind of rigorous level of information in order to start to work out where in particular the errors are arising.

I suppose the other point, more generally, about this conclusion is that, if I can refer you to the second bullet of paragraph 15, a lot of these exercises are now under way to be done by spring 2011. I think it's right; that will put the Department in a much better place, but I suppose we have obviously been assessing it on the performance we've seen to date.

Chair: Okay. Stephen, there's obviously been a fraudulent claim for heating in this room. We're being denied it; I don't know what you did. So Stephen and then I'm going to draw it to a close.

Q133 Stephen Barclay: Just following up on that to clarify something with Ruth Owen. This Committee recommended in 2008 that local offices with the highest error rate should be identified and targeted for more training and compliance checks. Is that now in place? Is that happening?

Ruth Owen: Only partially, I think, is my answer on that one. I think the previous report said we didn't have information down to site level; therefore we weren't able even to compare sites. We do now have that information. I have site-by-site management information. But it goes back to where the source of the error is. So, I could look at a particular benefit centre and say, "That looks like they're at the bottom of the league table," as it were, in comparison, but it might not actually be the error made in that site; it might have been earlier in the process—the contact centres or the job centres.

What I have done though, in terms of targeting, is we found through that root cause analysis that a large number of our errors are made in the job centres, and we didn't have a very thorough checking regime in the job centres. So I put a lot more resource into targeting checking in the job centres, by having done that analysis, the comparative analysis, of where the errors occur in the system, and made sure that we're now much tighter in checking that. So I think the answer to the question is in business model terms, yes, but not at site level, because it's not as straightforward to associate an error with a site.

Q134 Stephen Barclay: Right. Well, your answer actually opens up another issue, which is the independence of those doing the checking, because I think in Jobcentre Plus those doing the checking are based in the local offices. Do they report into local management?

Ruth Owen: In job centres they're done at regional levels. They're not actually in the site; they're one tier away.

Q135 Stephen Barclay: So the quality assurance teams are fully independent of line management?

Ruth Owen: Of line management, yes.

Q136 Stephen Barclay: Right, okay. Just to clarify in terms of my first question, where the Report, paragraph 2.8, says, "The data held centrally are not in a format to enable the management team to draw out statistically reliable information on the performance of each of its local offices," what you're saying is that issue has now been resolved.

Ruth Owen: Has now been resolved, yes.

Q137 Stephen Barclay: Okay, thank you. Coming back to Mr Oliver, this relates to something earlier. In terms of the lack of information that you would need, my question is whether you're using, in terms of tackling fraud, all the information Government holds. That particularly relates to the way local authorities are using that information and the interplay with SOCA. What sort of amount of fraud would you expect to be identified as a result of information coming from SOCA?

John Oliver: I don't think we can give you an answer on SOCA referrals through to fraud. I think we can do you a note on what we believe. What I can assure you is that we don't just work closely with SOCA. We work with most of the intelligence agencies and other Government Departments on sharing or on looking at best practice and between information and practice. If we're talking about fraud rather than official error, then we work very strongly with the likes of the National Fraud Authority and, indeed, the intelligence agencies on identifying serious fraud, which it sounds as if you're alluding to.

Q138 Stephen Barclay: No, I'm not actually; I think that's the misconception—that the information held by SOCA relates to really serious crime. Actually, SOCA sits on a massive database of information related to SARs reported by high­street banks.

John Oliver: Yes, I understand.

Q139 Stephen Barclay: There's a lot of information there. Now, the issue there is whether local authorities are contacting SOCA because they have a suspicion in the first place, or whether SOCA is just sitting on that information and local authorities have no incentive to get in touch with them in the first place, in which case is the Department looking at that in terms of its relationship?

John Oliver: We use SAR referrals. We also share that with local authorities. I can't pretend to give you the detail of that now. But I certainly know from our fraud investigative service that they do use SAR referrals as they come through. I don't know, Leigh, if you want to add to that.

Sir Leigh Lewis: We're quite a way from the subject matter of the report.

Q140 Stephen Barclay: It's 25% of the error rate, and I just wanted to use the opportunity, because what I'm driving at is not all banks provide the same information to SOCA. They're mandated in terms of certain SARs but not in terms of other things they're able to provide. The area in which they're not mandated relates specifically to certain data extracts they can gather on housing benefit and council benefit of accounts with high flows through them.

It just baffles me why SOCA sits on that information with no incentive for local authorities to get in touch with them, and the concern therefore is that one arm of Government is sitting on a lot of very useful information about potential fraud, which is exactly where investigators should be looking, but local authorities are not necessarily getting in touch with them in the first place.

Sir Leigh Lewis: Let me just say a couple of words on that. First of all, it picks up on the point that Anne McGuire was making. We have a single fraud investigation service for the entire Department, so it covers all of our business, all of our agencies. It has a very strong and good relationship with SOCA, and I think if there was a senior colleague here from SOCA, they would confirm that without equivocation.

We work very much hand in glove within the fraud investigation service, and I recognise the point you're making is not just about serious crime. But we have both, if you will, the mainstream fraud investigation service and what we call the organised fraud investigation service—I've seen that myself at very close quarters—which is seeking to tackle those people who are seeking to defraud us almost as an industry and to implement very major crimes. So we do work very closely with SOCA.

John Oliver: Just to add to that to reassure you. Our new strategic thinking is in the strategy that was published in October, where the Government announced the formation of a single fraud investigative service from 2013 that will pull together housing benefit, tax credit and DWP resource into the one service, and which will look at all the welfare and benefits.

Q141 Chair: Now, we're coming towards to the end. I just wanted really to question Ruth Owen and John Oliver. In answering the questions we've put to you today, every time we say there's a bit of action that you should have done, you're telling us you've done it—or you're doing it, or it'll be in by next April. Maybe I'll ask John Oliver first and then Ruth. Are you therefore confident that this 18 years of having to qualify the accounts because of the—let's just stick to the administrative error—£1.1 billion worth of administrative error, is that going to come down?

John Oliver: Chair, let me just say one thing. In the year that I've been in the role I've seen the biggest sea change of investment in this agenda from successive Governments that I think we've seen in the last decade. Some £450 million of investment, despite the fiscal climate, is a serious investment. We're looking at returns of up to £2 billion in the CSR period. That is significant investment. That translates to very serious investment in resource, where we're looking up to 1,400 resource looking just at data cleansing, cleaning out the dirty cases that we know are in our system, or proactively contacting customers about their change in circumstances.

Q142 Chair: Can you answer my question?

John Oliver: Yes.

Q143 Chair: By when?

John Oliver: Well, we're projecting that we'll begin to make returns on impact within the next financial year, but the problem that we've got with that is that the report lags; we've got 12 to 18 months before we see the results of our labours. But we're looking at significant improvements in our results within the next SR period.

Q144 Chair: So in 2011-12 we'll see an improvement, which, when those accounts get assessed in 2012-13, might lead to them not being qualified?

John Oliver: Well… oh, no.

Q145 Chair: Go on.

John Oliver: That's a different question.

Q146 Chair: Yes or no?

John Oliver: Could I answer on behalf of the National Audit Office on whether our accounts will ever be dequalified?

Q147 Chair: We know that the reason for the qualification is the fraud and error. We're talking here about the administrative error. You would hope by 2011-12, so by the time those were audited in 2012-13—

John Oliver: Let me answer you in one way. I think the Department is absolutely fully focused on getting those accounts dequalified and having an even more systematic approach to reducing official error.

Q148 Chair: Ruth, do you want to add to that?

Ruth Owen: All I would say is that I think since the NAO came to see us about this, which was 2008-09, I've definitely seen us do more. I think we're doing much more to address official error and you've heard today that there are more things we can yet do, so I would be very disappointed if that didn't flow through into the figures.

Q149 Chair: In 2011-12. Sir Leigh, really to ask you, what else does your successor have to do ensure that those accounts stop being qualified?

Sir Leigh Lewis: Succeed where I failed, because I set myself a challenge on the first day to lift the qualification of the accounts and I failed. That qualification remains there. Actually, joking aside, that is, I think, the biggest single disappointment I will take away with me on 31 December—that I did not achieve that, because I committed myself in every way to do it and we've not achieved it.

This is very complicated. For a long period it was an impossible goal, because the NAO under previous leadership set a 1% threshold, so we would have had to have brought 2.1% as the total level of fraud and error down to below 1% to escape qualification. That is simply impossible and I wrote the longest letter I have ever written to any Committee, to your predecessor, Chair, on 13 March 2008: seven pages as to my view on the qualification. It may be symptomatic, but I never got a reply, so it may just have been slightly too difficult a letter to respond to.

If I may say so, we've moved on a long way. I will not speak for the C&AG, who will speak for himself, but we are in a dialogue now that I believe has the real prospect of leading to the qualification being removed. I'll say one other thing, which is when you have had accounts qualified for 21 successive years it has lost any conceivable impact that you might want it to have. It has become a pure ritual that is of no use to anyone. We have to get to a point where there is a realistic prospect of this Department's accounts no longer being qualified.

Q150 Chair: Okay. Richard, do you want to ask your question?

  Amyas Morse: Can I just say something about this?

Q151 Chair: Oh, go on. Sorry.

Amyas Morse: You're all interested in my decision. Okay, just very briefly then Chair, because I'm aware that we've pressing other business.

Leigh and I have talked about this quite a bit. We have taken a somewhat different position to my predecessors, which is to say if it's possible for us—and we're working quite hard to try and do this—to look at what best practice would be in applying the benefits—not what best practice for the Department would be, but what objective outside best practice might be applied to a system of this complexity—if we can show that the Department's performance approaches that, then we think that probably would be enough for us to go forward.

Chair: Okay.

Amyas Morse: So that is a change of a significant nature. I don't mind saying we don't know if we can do it, but if we can that's where we'd go. So, there is some realistic prospect. It will require some significant further improvement from the Department as well as us developing something, which we hope to be able to do.

Q152 Mrs McGuire: Best practice in the banks did you say?

Chair: No, you mean internationally?

Amyas Morse: No, no. What we're going to do is take a number of institutions that do complex administrative tasks—

Q153 Mrs McGuire: And financial tasks?

Amyas Morse: —and we're going to break down what they do and find a way of carrying that across to apply it against the different things that the Department does. So it's not a question of just saying, "We'll take what the banks do and just apply that," because what they do is different, and in many cases somewhat simpler than what the Department's doing. It wouldn't be fair to apply it straightforwardly.

So if you can fragment that down into elements of information to be dealt with, elements of uncertainty to be accommodated and so forth, we may be able to do something in modelling terms that lets us come to a level there that would be reasonable; however well organised, nobody would get it below that level. That's what I'm trying to feel for. I don't want to base it on the Department's own performance; I don't think that's sufficiently objective for a regularity opinion. We're not talking true and fair; we're talking about regularity of use of public money.

Q154 Chair: Okay, now Sir Leigh, it's your last appearance.

Sir Leigh Lewis: My last hurrah.

Q155 Chair: I've had the pleasure of working with you both when you were at the old Department for Education and Employment and then in the DWP. Richard and I, before whom you've appeared innumerable times, want to ask: if you had a few words of wisdom to your successor, apart from getting these accounts out of qualification, what would they be? What would your advice to your successor be?

Sir Leigh Lewis: Can I say one thing about the discussion we've had? Perhaps it is a thing that you have the boldness to say on what is your last appearance with only nine working days of your Civil Service life ahead of you.

I think this has been a good, serious and sensible discussion, and it emphasises the point that there is more that we can do and should do on top of everything we have done, because we have not wanted, whatever else, for commitment. But madam Chair, I hope that as you come to write your Report on this hearing—I'll find it somewhere on your website—you will ask yourselves as well about the role of Parliament and successive Governments.

The NAO, in its report Dealing with the Complexity of the Benefits System in 2005, defined complexity as "complicated, involved, intricate; not easily analysed or disentangled," and concluded, "The benefits system fits this definition." Some of this has felt a bit like you're the passengers in the world's worst designed aeroplane and you're saying to the pilot, "It's a jolly bumpy ride," and the pilot's saying, "Give me a break; I'm keeping it in the air." With a system of this enormous complexity, I think that, while you can always do better, having official error at 0.7p in the pound is not a bad achievement, faced with a system of that complexity.

I think I only want to make one other reflection, if I may, and perhaps I'm just using the privilege of this occasion. I'd just like to say something about my Department's own staff, very finally to this Committee, because I've appeared in a sense on behalf of this Department's staff nearly 30 times. I've had the privilege to lead over 100,000 people for the last five years in the Department and before that in Jobcentre Plus. I've made a point of trying to go out and meet them and spend time with them every week. I haven't always achieved it on every week, but I've done it most weeks.

Overwhelmingly, I've met people who are hard-working, dedicated and committed, who don't earn a king's ransom and won't retire with a huge pension, but who are doing their very best, day in, day out, to serve the people of this country really well. Many of the jobs they do are tough, unrelenting and demanding. I'm very proud of what they do.

You should never let yourself as Permanent Secretary get irritated or cross, because it really doesn't do. The Permanent Secretary needs to be the one person who does calm, but if one thing does just occasionally irritate me, it's those caricatures that you continually read in parts of the media about Whitehall bunglers, pen-pushing bureaucrats and the like, and articles that almost assume that being a civil servant is somehow kind of synonymous with incompetence, laziness and the rest. I think that's a travesty. It's a travesty of what I've seen in the Department I've led and I think the commentators and journalists who write in those terms should think very hard before they do so again.

So, madam Chair, that's it; I'm not going to perhaps miss receiving those wonderfully, beautifully worded invitations to attend the PAC in future, but I think together we've been trying to do something quite important, and that's improve the government of this country.

Chair: Well, I think from my own perspective, and certainly of those of us who've worked with you, you will be sorely missed. I think your contribution to public service has been terrific, so thank you very much for that.

Q156 Mr Bacon: May I come in and just prolong your agony for a minute or two? I add my endorsement to what you've said about your civil servants. My own experience of Jobcentre Plus locally has been of very serious and committed public servants who know what they're about, know why they're there, know why they're trying to do what they're trying to do and are coping with an extraordinarily difficult and complicated system. There is, of course, the underlying question of why it's complex, who made it complex and whether it's Ministers being insufficiently challenged when they come up with silly ideas or whether it's just the accretion over many, many years of trying to do the right thing.

It's very striking, looking at both Ruth Owen's and John Oliver's CVs, that you have sitting next to you two people who have come up from the very bottom: John Oliver as a direct entrant Executive Officer; and Ruth Owen, who started in the new model DSS office as a National Insurance inspector.

From listening to both of them, you get a deep sense of people who know what it's like on the inside from the bottom, and who know what it's like to have been at the bottom. I noticed, Ms Owen, you were in the design authority for the Jobcentre Plus—with Lesley Strathie, presumably, who also started at the bottom. That to me is a very important component of getting it right—having people who understand things from the bottom.

You now face a very, very big task; it won't be one for you, Sir Leigh, because you're leaving the Department, sadly. May I say, in parenthesis, how much I've enjoyed listening to you over the years. You've always managed to be both courteous and certainly give the impression of being tremendously candid, without, I think, ever saying anything you intended not to say. In some ways, you're the model of a modern Permanent Secretary.

Chair: I'm not sure about that.

Q157 Mr Bacon: The challenge, which you've identified, is to get rid of some of this complexity by radically re­engineering the whole benefits system, and the apparent answer is the universal credit. In the desktop research I've been doing while sitting here I've come across prices for that varying from £2 billion to £3 billion for the upfront costs of implementation.

Could you just say a little bit about that? Who's in charge? Who's the senior responsible owner? What is the proposed budget for implementation? Who's the computer contractor? How long is it expected to take? Why will it not fail in the way that other projects have often failed? What safeguards have been put in place to make sure that it comes in on budget and on time in the way that you've said some other projects have done? It is going to be tremendously important for the transformation of your Department, isn't it?

Sir Leigh Lewis: It most certainly is. Let me just give you what I hope is some reassurance, almost picking up on points you were making about John's CV and Ruth's CV. The person who is the senior responsible owner, and who has just been moved into that role, and who Anne McGuire will know very well from her time in the Department, is Terry Moran, who until very recently was the Chief Executive of the Pension, Disability and Carers Service.

He's one of the civil servants, I think anywhere in Whitehall, with the most delivery experience. He started as a 16-year­old; he started as a 16­year old in the Department and has worked his way up through every branch of its operations to the level of Director General. I think that does say something about the importance we attach to direct delivery experience.

Now, no one can ever say that a programme of this kind is without risk; of course it has risk in it, but absolutely we're putting in place a really strong governance structure. It is operating and will operate to a Permanent Secretary overseeing the process, with very, very strong ministerial agreement, and the funding was announced as part of the overall spending review. So, of course this is going to be a huge challenge for my successor and my Department, but I am genuinely confident of our ability to deliver. I am genuinely confident of our ability to deliver.

Q158 Chair: Good. Thank you very much indeed and all the best for the future.

Sir Leigh Lewis: Thank you.


 
previous page contents

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2011
Prepared 10 March 2011