UNCORRECTED TRANSCRIPT OF ORAL EVIDENCE To be published as HC 1025-i

HOUSE OF COMMONS

ORAL EVIDENCE

TAKEN BEFORE THE

PUBLIC ACCOUNTS COMMITTEE

TAX CREDITS FRAUD AND ERROR

WEDNESDAY 6 MARCH 2013

SUE ROYSTON

LIN HOMER and NICK LODGE

Evidence heard in Public

Questions 1 - 162

USE OF THE TRANSCRIPT

1.    

This is an uncorrected transcript of evidence taken in public and reported to the House. The transcript has been placed on the internet on the authority of the Committee, and copies have been made available by the Vote Office for the use of Members and others.

2.

Any public use of, or reference to, the contents should make clear that neither witnesses nor Members have had the opportunity to correct the record. The transcript is not yet an approved formal record of these proceedings.

3.

Members who receive this for the purpose of correcting questions addressed by them to witnesses are asked to send corrections to the Committee Assistant.

4.

Prospective witnesses may receive this in preparation for any written or oral evidence they may in due course give to the Committee.

Oral Evidence

Taken before the Public Accounts Committee

on Wednesday 6 March 2013

Members present:

Margaret Hodge (Chair)

Mr Richard Bacon

Chris Heaton-Harris

Meg Hillier

Mr Stewart Jackson

Fiona Mactaggart

Austin Mitchell

Amyas Morse, Comptroller and Auditor General, Gabrielle Cohen, Assistant Auditor General, NAO, John Thorpe, Director, NAO and Marius Gallaher, Alternate Treasury Officer of Accounts, were in attendance.

REPORT BY THE COMPTROLLER AND AUDITOR GENERAL

Tackling tax credits error and fraud (HC 891)

Examination of Witness

Witness: Sue Royston, Policy Officer, Welfare Policy Team, Citizens Advice, gave evidence.

Q1 Chair: Welcome, Sue. Thank you very much indeed for coming. We are a little short on numbers today, because some of our members are giving evidence to another Select Committee and a couple are abroad. I am grateful to you for coming. The reason for you being here is that all too often we get seduced by just looking at systems and bureaucracy. You are on the front line, dealing with many of the clients for whom the system is not working at present. In our surgeries we get quite a lot of people who have tax credit problems. It would be helpful to get your perspective on what you think the issues are that create problems, and what you think should improve in the system.

Another preface is that this is megabucks. A lot of money is being lost. Clearly, if the Government do not get the money that they are losing through fraud and error, they will have to cut other services to cut the deficit, so we all have an interest in ensuring that fraud and error are minimised. Over to you. Say whatever you feel like, and then people will ask you questions.

Sue Royston: We recognise how important it is. We believe there is not enough emphasis on preventing overpayments. If there was good quality information, exploration, and advice when somebody is making a claim or reporting a change, we believe that a lot of the problems we see could be prevented. The information is often inaccurate. The disability element was a good example of that.

Q2 Chair: Inaccurate from whom?

Sue Royston: Sorry. From the tax credit helpline.

Q3 Chair: Not from the claimant?

Sue Royston: It may be inaccurate from the claimant, but what we are concerned about is that the information is inaccurate.

Q4 Chair: Just give me an example of that sort of thing.

Sue Royston: A couple of years ago, there was an investigation into people who were claiming the disability element of tax credits, and a lot of people had their money stopped. We discovered that a lot of them had had their money stopped incorrectly. We did a mystery shopping of the helpline and discovered huge amounts of incorrect information. Basically, most people on the helpline and within the Tax Credit Office thought that the only route through which you could get the disability element was by claiming DLA. They did not know about the sickness benefit route. That cost a lot of money, because the tax credit office had to go back and relook at the claims they had stopped, and many of the claims were incorrect. We believe there should be more accurate information at the beginning of the claim. That is one extreme example.

There could be more exploration at the beginning. Self-employment is a good example. We see people who made a claim two or three years ago and at the time said what they were going to do, what their business was and what they were going to be working on. Two years down the line, the Tax Credit Office has said, "You are not working in expectation of payment."

We feel there should have been a much firmer gateway at the beginning of the claim. We would suggest something like people having to make a business case, say what they are going to do, and put forward the steps that they are going to take. The Tax Credit Office would decide at that point whether that is in expectation of payment, rather than retrospectively deciding, because it did not actually produce payment, that it was not in expectation of payment.

There are a number of examples like that, where not enough exploration and guidance is going on at the beginning. Another example is lone parents. We recognise that the Tax Credit Office sees a lot of examples where it feels that the person just has not told it. However, the lone parents we see have actually said they were separating and could have given all the information about that separation at the time. It was not explored in enough detail. Two years down the line, they are being told, "You aren’t really a lone parent; you are a couple."

Q5 Chair: Explain to me how that happens. Either you are or you aren’t. It is a big area where they think a lot of fraud and error goes on.

Sue Royston: It is a very big area. It is not a straightforward or simple situation. People’s lives are very complex when they are separating, especially when there are children involved and a mortgage being paid, and when they have been together for some years. I hope you have the document I sent you.

Q6 Chair: Yes, we have.

Sue Royston: As the situation is so complex, I set out this flow chart. It shows the steps that are gone through and all the evidence we have received about the problems this is throwing up.

Q7 Chair: Can I just stop you for a little bit? The Report says that, in a sense, HMRC agreed with you and put a lot of effort into prevention. It was actually pretty ineffective, and it did not get the money. One of the reasons why it has not met its target on fraud and error is that it did not get about a third of the money-John will correct me if I am wrong-that it anticipated it would from that early prevention work.

John Thorpe: Yes, the numbers were restated. Initially, the estimate for 2010-11 was about £1.4 billion, and on restatement that was £480 million.

Q8 Chair: Was that all early prevention stuff, or was that everything?

John Thorpe: That was a mixture of interventions around prevention, and looking at the case load and correcting errors in claims that were already in the system.

Q9 Chair: We all think that prevention is better than cure, and I would say the same for my case load, because we all get a lot of tax credit casework. In terms of limited resources from HMRC, it did not look like a massive bang for its bucks.

Sue Royston: I do not think that I can really comment on that. What we can say is that people are being put in a lot of hardship as a result of this process. Retrospective decisions are being made that should have been made at the time of the claim, or at the time of the reporting of circumstances. In many cases-I recognise that the Tax Credit Office is dealing with large numbers of people in the undisclosed partner project-not enough care is taken over the decisions. I am getting a great deal of evidence from advisers across the country about people-I am sure that MPs have seen this in their surgeries-who we believe are genuinely lone parents, and who are having to live for months on end on income support and child benefit because they are no longer getting tax credits. That is a great deal of hardship. As for those sorts of retrospective decisions and stopping somebody’s benefit, we do not think that is a fair position to put people in.

Q10 Austin Mitchell: I get the impression, from my experience, that a lot of the problems are because the machinery is so big, bureaucratic and cumbersome. I notice that you say, in this analysis of the evidence, that in one case a claimant provided two supporting letters from neighbours who are serving policemen, but that was all rejected and ignored.

Sue Royston: Yes.

Q11 Austin Mitchell: So is it the case that local evidence is being ignored, and is that because of the scale of the bureaucracy?

Sue Royston: I think so, yes. We are certainly getting a lot of reports of local evidence being ignored, or decisions being made and then just not looked at carefully, perhaps because of the scale of the problem. I accept that, but it causes enormous hardship, because there is such a backlog of appeals. It means people have to live without tax credits for long periods.

Q12 Austin Mitchell: If supporting evidence is being ignored, what about snitches who want to grass the neighbour? Is there a lot of that? Is that ignored?

Sue Royston: I am not aware of that; I do not know whether that is reported. I was looking at information from the benefit hotline, and I know that of every 75 calls to the benefit hotline, only one actually proves to be fraudulent. I think in a lot of cases people are not aware of their neighbours situations, but there is local evidence that could corroborate what the client is saying.

Austin Mitchell: Yes.

Q13 Fiona Mactaggart: Has the number of appeals from lone parents changed since 2009, when the policy changed? Has it gone up or down?

Sue Royston: In the last year, over the undisclosed partner project, we have seen a huge increase. It has always been a small area, so we are not talking huge, in terms of statistics, but by looking at separate quarters, we think we have had an increase of about 700 clients nationally, quarter to quarter.

Q14 Fiona Mactaggart: What is that as a proportion?

Sue Royston: Just under 3,000 extra clients. That is looking around the area of compliance, cohabitation and appeals for tax credits.

Q15 Fiona Mactaggart: If those appeals succeed, how long does it take to get a successful outcome?

Sue Royston: This arose as an issue about a year ago in Liverpool, because that is where the pilot project set off. I do not know the number of appeals we have got waiting nationally, but so far only one appeal has been heard-that was in December-and the client won that case. That was a client who came much earlier in the year because she was in Liverpool, where the pilot project was. Clients are getting letters saying, "Do not contact us for six or eight months."

Q16 Fiona Mactaggart: So in effect, for those who are correct, HMRC is borrowing your clients’ money for months.

Sue Royston: Yes, and it is causing enormous distress. We have seen one client admitted to hospital with suspected malnutrition. In a couple of weeks’ time we have got a client going into a possession hearing about her council flat, because she works part time and has three children, and she has had to use the money she has for rent to feed her children. Advisers are telling me that they have to keep on issuing food vouchers. It is causing enormous distress.

Chair: This is the appeal process?

Fiona Mactaggart: I am diverting slightly, but I think it is important that we recognise that the slowness of the appeal process means that there is a fundamental injustice in how this works, as well as the issue of fraud and error. That is all I wanted to highlight.

Q17 Meg Hillier: I am looking at your very helpful evidence trail about the appeal deadlines. The letter from HMRC saying, "Do not contact us for six months" reminds me of my days working with the UK Border Agency. Is this something recent? Has that deadline increased? Or has that always been the standard letter?

Sue Royston: No, it is because such a lot of appeals have come in.

Q18 Meg Hillier: How long has it been six months?

Sue Royston: I am not sure about that, but I do know that they have had to put a lot more people into the appeals process, and they are hoping that it will hurry up. We feel, again, that it should have been seen that if you are doing something on this scale, and sifting through in this way, there will be a lot of appeals and this sort of thing will happen.

Q19 Chair: Why do you think the errors are going up? You have looked at the Report, haven’t you?

Sue Royston: Yes.

Q20 Chair: You saw the categories where the errors are going up: people reporting about partners, changes in children’s circumstances and a third one that I cannot remember. Why is that happening?

Sue Royston: I think one of the issues about hours of work is that more people are getting zero-hours contracts. Obviously, there are very strict hours rules for tax credits, and you fall out of entitlement if you work-

Chair: Less than 16 hours.

Sue Royston: Less than 16 in some cases, less than 24 in others, and less than 30 in others. That is a real issue, because if you are getting a different number of hours each week and it is averaged, you can very easily drop below the average without realising it. That would be one of the reasons why that is happening.

What we are seeing is not so much people not reporting things; it is decisions being looked at again and judged a different way. Just to give you an extreme example, one of our clients is being pursued simultaneously by a compliance officer who says she is in a couple, and a compliance officer who says she is a single person. Both say that she should have claimed a different way. The bureau is trying to persuade each compliance officer. That is a measure of the different ways that compliance officers are looking at and re-judging situations. It puts the client in a very vulnerable situation.

Q21 Austin Mitchell: There is a television programme there, in which the two compliance officers eventually meet and get married, but that is not what my question is about. Is it your impression-it is certainly mine-that most of the problems with the system that have arisen, and arise constantly, for our constituents are not due to deliberate fraud, and an attempt to milk the system, but fluctuations in earnings, caused particularly by the increase in part-time working on low wages, and bureaucratic delays in registering those? The system cannot keep up with the changes in earnings, so people get penalised, but it is not their fault.

Sue Royston: Yes; certainly, in terms of overpayments, that is true.

Q22 Austin Mitchell: So that is your impression, too?

Sue Royston: Yes. As for the other areas that the National Audit Office identified, there was the undisclosed partner, and as we have said, we feel that that is about retrospective decision making-not in all cases, though; people do not come to us if they want to defraud the system. They are not going to put themselves through talking to somebody else.

Apart from the undisclosed partner project, there are a number of other areas where the Tax Credit Office has been very slow to provide information. To give a very clear example, it has identified that children leaving school at 16 is a problem area. The Child Benefit Office, since time immemorial, has sent a form when a child is reaching 16, asking, "Is your child staying on at school?" That is run by HMRC, and the Tax Credit Office is run by HMRC. Clients have had a form from the Child Benefit Office, but not from the Tax Credit Office. They naturally assume that the Tax Credit Office knows about the system. There are all sorts of things such as run-ons of child benefit. If the child is out of work and the adult still has to support them at home, the adult thinks, "Okay, the office knows the situation; I have declared it on the child benefit form." I think the Tax Credit Office has moved on that issue, but there are far too many situations where it has been too slow to provide the right sort of information and scenarios.

Q23 Chair: We are about to move on, but do add anything that you have not had the chance to say. With universal credit, you will be dependent on yet another player in the field: the employer. You can have an individual getting it wrong, the tax credit world getting it wrong and, now, the employer getting it wrong. Have you thought through the implications of all that, and how you think that can be best managed?

Sue Royston: Yes. There are a number of concerns. With universal credit, if RTI works, it does have the potential to be helpful, because you then have real-time information.

Q24 Chair: My example is a pub owner who employs a lone parent for a few hours a week, and the hours change every week. The info might well be wrong. Not deliberately-just by mistake-he or she will fill in the wrong bit of the form.

Sue Royston: Yes. We are very concerned, because at the beginning of the tax credit system, money was paid through working tax credit, using that information. They had to stop that, because what was seen to happen was that a lot of employers were sending the wrong information. The client is then in a very vulnerable situation. They do not want to have an argument with their employer, because they do not want to lose their job. We saw clients who lost tax credits because they were not prepared to argue with their employer about the information that they had given. In the example you give, that is clearly a worry.

John Thorpe: Perhaps the Department can answer this later on, but on the RTI situation, my understanding is that there is a control in the process that reconciles the payment that the individual receives with the information submitted through the RTI process. So if that process works correctly, the information that the employer submits should be in line with the individual’s pay details and what they have actually received.

Chair: Say that again, because I was a bit distracted.

John Thorpe: The Department can probably answer this, but it is my understanding that, under RTI, there is a control in the process that reconciles what has flowed through the payments process with the actual earnings and tax deductions, so that information should be in line with what the individual is receiving. There is a control there.

Chair: Okay, so it shouldn’t be a problem. Is there anything else you want to add before we move on to the main session-anything you feel is absolutely key?

Q25 Austin Mitchell: Can I ask something, before you add what you want to add? I have had cases where a couple-sometimes married, sometimes not-split up and the bloke moves out and takes the tax credit with him, leaving the woman to pay the debts. Why does that happen, and is it a common occurrence in your experience?

Sue Royston: It can happen. The child tax credit would normally be paid to the main carer.

Q26 Austin Mitchell: But it doesn’t necessarily stay with the main carer.

Sue Royston: I haven’t come across a situation where the child tax credit has gone, to be honest. I am not saying it does not happen.

Q27 Austin Mitchell: So it would take fraud for the deserting carer to claim that he was the carer?

Sue Royston: It probably would, yes.

Q28 Chair: Anything else?

Sue Royston: I would just say that we are really concerned that universal credit gives guidance up front about things. A lot of these situations are very complicated. It isn’t straightforward when somebody separates or when somebody becomes a partner. Scenarios guidance and leaflets about the situation would be very helpful. There should be much more information for the clients.

Chair: I agree. You would help my caseload. Thank you very much indeed. We are very grateful.

Examination of Witnesses

Witnesses: Lin Homer, Chief Executive and Permanent Secretary, Her Majesty’s Revenue and Customs, and Nick Lodge, Director General Benefits and Credits, Her Majesty’s Revenue and Customs, gave evidence.

Q29 Chair: Right, Lin, welcome. We are seeing you today and tomorrow. I think you are going to need a gin and tonic after the slot-although there is no gin down that end of the table.

I read this Report, and I am quite shocked-more shocked, probably, than the NAO were when they wrote it. This is a mega-problem, it seems to me. HMRC has been trying to tackle fraud and error since 2008 as far as I can tell. I am hugely disappointed at the lack of progress. Let me unpick that a bit. For example, you do not know what is error and what is fraud, and, within error, you don’t know what is claimant error and what is error by the Tax Credit Office, do you?

Lin Homer: We think we know some of those things. We will happily try to unpack them a little.

Q30 Chair: Sorry to interrupt, but they are not there in the Report. When we do DWP fraud and error, which is also a big problem-although it is down at 2.9%, not 8.1%-we get a breakdown that talks about administrative error and client error. On this, what we have is a massive fraud and error without any of that breakdown.

Lin Homer: Shall I try to respond to that and see if we can help? Much of what the Report says-and, indeed, what Sue said to you-we recognise and accept. This is a very complex system. We believe that the majority of what we are seeing in the overpayment area is error rather than fraud-between 70% and 80%. I can unpick that a bit more. We share your disappointment that we have not made more progress in reducing that as a percentage. We set ourselves a target in 2009 for 2011 to reduce it to 5%, and we missed that target. We shall share with you during the hearing what we are doing.

Chair: Okay.

Lin Homer: There is an in-built element to the way in which tax credits are undertaken that almost builds in the risk of error. We make a provisional payment and, at the end of the year, claimants are required to confirm that what they thought would happen has and, in a small number of cases-thankfully, going down-we have underpaid, but in a fairly static number of cases, a significant number, we have overpaid-one in five of the claims.

The system was designed to be responsive, partly to avoid the problem that, if you make a snapshot payment and things get worse for claimants, they are left unprotected. But, of course, once you make it responsive, if their situation improves, they finish the year owing you money.

The question that Sue brought out is whether people understand when their circumstances have changed enough to tell us. We would say that that is a real problem. We have not talked to you about them yet, but we have done some things to try to help claimants through what we would accept is a challenging and complex process.

Q31 Chair: Yes, I think that it is far too complicated, but lots of us think that. So 70% is fraud.

Lin Homer: No, error.

Q32 Chair: Of that 70%, how much is claimant error and how much is Tax Credit Office error?

Lin Homer: Can I ask Nick to unpick that a bit for you? You have met Nick Lodge before. He now heads up this area for me. He came with me when we did the accounts.

Nick Lodge: To break down the total figure for error and fraud, it was 8.1% for 2010-11, that equates to £2.3 billion. Of that £2.3 billion, £1.6 billion is error and the rest we would categorise as fraud. Very little of it is HMRC error because of the nature of the system, so about £40 million of that £1.6 billion is HMRC error.

Q33 Chair: £40 million.

Nick Lodge: £40 million-about 2.5%. This is all change of circumstances.

Q34 Mr Bacon: You are saying that, of the £1.6 billion, £40 million is-

Nick Lodge: About 2.5%.

Q35 Mr Bacon: So 97.5% is claimant error, so that leaves £700 million as fraud.

Nick Lodge: Yes. That is correct.

Chair: I want to unpick these figures a bit more.

Q36 Fiona Mactaggart: When people appeal, what percentage of appeals succeed and what percentage of appeals represented by the CAB, for example, succeed?

Nick Lodge: The appeals process is that, first, the claimant would appeal to HMRC. We try to settle those appeals with the claimants before it goes to an independent tribunal, which is the next stage.

Q37 Chair: Except you tell them not to write to you for six months.

Nick Lodge: Well, we have actually now written to all claimants who have appealed as a result of one of our checks and either informed them of the outcome of their appeal or asked for more information from them. Generally, that can be very helpful to the claimant if they have submitted an appeal with no or little information, but that does not give us, or the tribunal later, any grounds on which to overturn the original decision that we made. We try to be helpful and ask for additional information.

To come back to your question about the number of appeals upheld versus the number that are not, to put the appeals numbers in context, for 2011-12, when we would have carried out approaching 1.5 million checks, we received appeals in 0.6 of 1% of cases.

Q38 Chair: What percentage?

Nick Lodge: They are all absolutely important.

Q39 Chair: They are big numbers.

Nick Lodge: That was about 8,000. It is about 60:40 in favour of HMRC, roughly. For the cases that Sue was talking about, which are particularly to do with undeclared partner checks, few have actually reached a tribunal-about 60-and, of those, 80% have been upheld in HMRC’s favour.

Q40 Chair: Let me unpick these figures a bit more. Often in our surgeries these are the most terrible cases that come to us. It tends to be lone parents and people who are really struggling to play the system right and true, and they suddenly are confronted with a bill that completely throws them. Sue gave one case. Often they are people who are trying to work, who have kids and who think they have got their tax credit right. Years later, someone tells them that it is wrong and they have thousands of pounds of debt, which they can never cope with.

Let me unpick those figures, because there is that human bit on one side, but on the other side you have a Government who are trying to cut the deficit, and your ability to cut fraud and error is poor. Let me go to a response that David Gauke gave, looking at tax credit overpayments, which is another bit of this world where you know that you have overpaid, rather than you do not know-that was how it was defined to me. These figures are shocking. You look at remissions, and in 2007-08 you remitted £240 million. By 2010-11, that was up to £600 million. Last year, for 2011-12, it was about £1.6 billion or £1.7 billion. That is a massive amount of overpayments that you know about, and you remitted. That means that you shoved them in a box where you never expect to ever collect them, but you are too scared to write them off. Your write-offs were tiny. That is megabucks for the Treasury and the taxpayer.

Lin Homer: When we were talking to you about our end of year accounts, we said that the level of debt generally in HMRC and the level of debt on tax credits is significant and challenging. You summarised it quite well when you were having the discussion with Sue. On the one hand we have got individual families and their case and whether the children and family are getting everything that they are entitled to, but on the other hand we have an annual figure of £1.6 billion of overpayments. That, relative to the country’s austere position, is a big amount of money. All I would say is that it is partly the nature of this system that we register and make a claim. A great deal of work has been done in recent years to make the claims accurate at the time that they are made.

We think that only 20% of the error we are looking at is from inadequate decisions at the beginning. Some 70% of the error comes at the change of circumstance moment. The system is designed, in a way, to run into lag. I write the letters that many of you receive in answer to your constituency queries. At the end of the year, when the claimant gives us up-to-date information and we rectify their claim, we will always go back and tell them that they have been overpaid, but the nature of the system lags how quickly we take the money back. You get overlapping consequences over years, and that is very difficult for the claimants and for us. One of the things that Nick and I have been talking about is whether we can make more rapid reclaims when inadvertent overpayments have been made, so that people do not get themselves into complex year-on-year debt.

Q41 Chair: The system was introduced in 2002-03. I accept that it is a ridiculously complex system that is difficult to administer and I accept that there are easier ways of helping people in need, but we have got this system. We are now nine years on and this is 2011-12. The figures are shocking. Let me just go on, and I will hand over. I was pretty appalled.

The only way that you will be able to get a grip on this massive loss to the taxpayer is through timely information. Then I read to my horror that you do not even ruddy well reconcile your error-let’s take the error part-until 15 months after the year end. That is plain bonkers. I do not know why you get to that point. As I read it, I thought that it may be that the best is the enemy of the good and you will not put out a figure until it is 100% accurate, but you must know what you are dealing with much earlier on. If you do not, you jolly well should, and then you would be able to tackle it. Then we would not have these scary figures that we have to deal with here.

Lin Homer: On the point that you make about more timely interventions-

Chair: Information-just data.

Lin Homer: Whatever point you like, Nick could give you a list of the things that we are trying to change to get that accuracy. On the general point of the lag, this is the system. So the claimant is not, in the main, required to tell us about change of circumstance until the end of the year. It is not that we choose not to reconcile. We-

Q42 Chair: But it takes you 15 months from the end of year to have the data on overpayment and underpayment-15 months. So you will know about April 2012 in June 2013.

Lin Homer: This is because this is based on the tax system, so we begin a claim in a tax year. At the end of the tax year, the claimant gives us information. We do work all the way through-Nick’s team are making assessments-but the retrospection tells us that we do not always get those as accurately as we would like.

I guess that this is the point where I do the segue into real-time information, which I know that we will talk to you about several times this year. Although we have got a lot of work to get people settled into RTI, we believe that-both for claimants and for us-real-time information will be a real boon in the area of tax credits. It will give us a source of real-time information about their earnings, which will allow us to check and validate major shifts much more quickly-

Q43 Chair: Honestly, Lin, I am happy to hear that, but if you are lucky-we have heard all sorts of rumours in the press-we will get there by 2018, 2019 or 2020. You can see that that will be a better route. Now, we have had this system in since ’03, so we are probably a little over halfway through its life and we cannot wait. You have got to find-

Lin Homer: No, and that is why we have been doing things-

Chair: We cannot wait. We see this level of remission-I hate that word, because it means that you are never going to get it. I would much rather that you were honest and wrote it off-and a huge level of fraud and error, which is just awful.

Q44 Mr Jackson: A bugbear of mine is about data-admittedly on a narrow area. On page 10, recommendation d says: "HMRC should assess how it can make better use of available data and analytical techniques." If you go back to page 8, in the summary it talks about children. One of my bugbears is about child tax credits for EU citizens whose children do not live in the UK. It is intensely irritating for people to hear-not you per se-when you are extolling the virtues of Government policy on top-rate tax child benefit and saying how it is all going swimmingly when they know that money is being remitted for families whose children do not live in the UK. That would be bad enough, but you do not know how many there are. Routinely, we ask Treasury Ministers what the estimate is of how many, or the cash sum, and invariably we get the response that it is not cost-effective for the public purse to find out that information.

My question to you is: is that really the case? Is it really beyond the bailiwick, the locus of HMRC to collect that data? There may be legal issues about whether we have to pay that money, but if we do not even collect the data, how can we make an estimate specifically of the potential fraud that might be perpetrated against the taxpayer?

Lin Homer: Can I ask Nick to answer that question?

Mr Jackson: Either of you is fine.

Nick Lodge: We do have numbers on the number of child tax credit claims that relate to migrants whose children live abroad. I cannot remember the exact number, but broadly the number of claims in the system for that circumstance, out of the roughly 5.5 to 6 million families in the tax credit system, is 4,000. The number of children in respect of whom payments are being made is about 7,000. Those are the latest figures as at December.

Q45 Mr Jackson: I think those figures are utter rubbish. There were 34,500 national insurance numbers issued in the Peterborough unitary authority from 2004 to 2012. That is 34,500 and most of those people are young people. A huge number of them are children and you are saying that it is 4,000 for the whole of the United Kingdom. I simply do not believe those figures. Even if they were true, why aren’t Treasury Ministers routinely revealing those figures to Members of Parliament when they ask questions, time after time and year after year?

Lin Homer: There is a difference between the figures we remit for children outside the country, which is what you asked, versus the figures we are paying for children inside the country. Secondly, as we probably say in a number of Committees, there is a difference between us sharing our published information at year end which will tend to update annually and trawling our information every time a PQ is asked. I would never knowingly remove that figure from you, but I would probably resist updating it every two or three weeks if you asked me again. That would then require us to manipulate the data.

Q46 Mr Jackson: Would that it were that nuanced. David Gauke says that it is too expensive for the taxpayer. It is clearly not too expensive to keep remitting money to people whose children don’t live in the UK. That is fine. But it is too expensive to facilitate transparency and openness to an elected Member of Parliament. It is not just me. Anne Main, the Member for St Albans, has been trying for years and years. Would that they were that nuanced, even once a year to say that this is the annual figure. But we are constantly told that it is not cost-effective to quote this data. It goes against the principles of openness and transparency. I don’t believe those figures and I think you need to look at this issue to reassure people that you have value for money at the forefront of what you do.

Lin Homer: Well, Chair, we have talked several times about us getting a better pack of published information. I am quite happy to take that away as a piece of information that we should make sure is there in the annual report somewhere. It seems valid to me.

Q47 Chris Heaton-Harris: Can we get a note with the exact figures as well?

Chair: Did you hear that? Can we have a note? We are being a bit bolshie about notes. We want them on time, please.

Lin Homer: I am usually very good at that but I will make sure we are.

Q48 Austin Mitchell: Do you check up that the kids actually exist?

Lin Homer: We have quite a complex system. One of the things we get quite a lot of letters from MPs about is that we take rather longer on these claims than on the in-country claims for just that purpose.

Q49 Austin Mitchell: Yes, but if I come in as a loud-mouth vulgarian, I mean Bulgarian-

Mr Bacon: Not, let’s face it, difficult for you to do.

Austin Mitchell: Louise Mensch did say that I was that. I think she got the V instead of the B. If I say that I have 16 kids, how do you know?

Nick Lodge: In these kinds of cases we check with the host country, the competent authority-

Q50 Austin Mitchell: But how do you know that there is not a perambulating set of kids who are pushed in for any claimant?

Mr Bacon: Single transferable kids-like in the CAP, with cows going across the borders between Greece and other countries.

Nick Lodge: We check the composition of the household with the competent authority. This is what takes time. So these international claims generally can take two or three months or perhaps longer to process because we run all of these checks with the host country.

Q51 Austin Mitchell: But there must be an enormous amount of fraud in this area. Why can’t you quantify it?

Lin Homer: Well, Chair, I would just challenge Mr Jackson. The figures we gave you are the sums we pay out. So those are relatively small in absolute numbers and in percentage numbers they are tiny. We, of course, also answer claims in relation to British parents similarly claiming child benefit in other European countries.

Q52 Austin Mitchell: But other European counties don’t pay this so why should we have to pay it to people from other European countries that don’t have that system?

Lin Homer: Well, others do.

Mr Jackson: We are not anti-European.

Mr Bacon: How could we be? We live here.

Q53 Chair: Can I just get back to what you said earlier? You said something about people who were-

Lin Homer: We answer every year a certain number of claims from Britons who are-

Q54 Chair: Who are living in Europe?

Lin Homer: Working in Europe and whose children are still living in Britain, choosing to claim the European benefit for their children. Mr Jackson made the point. European law is very clear on this, unless it has changed-

Mr Jackson: It is not very clear-[Interruption.] May I just finish my point? It is not very clear. Another issue is that the DWP has never properly looked at what a habitual residence test actually means, so it is not clear at all. Perish the thought that you should ever wish to test this in order to defend value for and the interests of taxpayers and my constituents.

Fiona Mactaggart: It is clear-

Mr Jackson: It is not clear. That is the whole point, and why this is a moot point.

Austin Mitchell: May I move on to another area, loth though I am to do so?

Q55 Mr Bacon: I would like to stay on this area for a minute. You mentioned the 7,000 children involved: what is the value of payments made in a year?

Nick Lodge: The value of payments is not something that we quantified-

Q56 Mr Bacon: What? Surely you just add them up? If you know there are 7,000, you must know what you are paying to each one.

Lin Homer: That will vary all the time. It is children going-

Q57 Mr Bacon: How can you possibly account to Parliament for how you are spending public money if you cannot begin to answer this sort of question?

Lin Homer: I would challenge that, Mr Bacon. We have given you the figures-

Q58 Mr Bacon: Which figures have you given?

Lin Homer: We have given Mr Jackson the figures about the number of families-

Q59 Mr Bacon: Four thousand families and 7,000 children. I want to know what that is costing my constituents and others, who are taxpayers.

Lin Homer: If we can write you a note, we will try to give you-

Q60 Mr Bacon: I do not want a note; I want to know now what this is costing, and it sounds like you cannot tell me, which surprises me.

Lin Homer: Well, I do not think it should, because-

Q61 Mr Bacon: Given HMRC’s record, I suppose that is fair.

Lin Homer: No, that is unfair, Mr Bacon. We have come to discuss error and fraud. All I would want to do, again, is to put the scale of what you are talking about in the context of this Report.

Q62 Mr Bacon: Yes, you did that quite well earlier. You said that it is absolutely tiny in the context of the total amount paid. That is true, of course: if you set up a system that is paying out many, many, many billions of pounds, then a mere few million by comparison is relatively, statistically, a very small amount, but it is still real money and it still has to be earned by our constituents who pay tax. Money is handed over to the Government, who hand it out but know not how much-you cannot say how much-and that seems strange.

Lin Homer: We have not come armed to talk about every subset-

Chair: Can you put it in a note?

Q63 Mr Jackson: But you are going to have 29 million people who are eligible to travel from Romania and Bulgaria on 1 January 2014-

Chair: They are not going to, Stewart, you know that.

Mr Jackson: If I may finish, you are going to have that situation, and it is quite discombobulating, if I may say so, that you do not know the cash sum. What if that 7,000 goes up to 70,000? Will you nonchalantly tell us that you do not have the cash sum: "I will write you a little billet doux later about it"? You should have the cash amount now: it is only 7,000 children. We are talking about error and fraud; you should have come with that information. It is a pertinent political issue that has caused concern to Members of Parliament for many years.

Lin Homer: I am very happy to write. I am also content to say that we bear down very heavily on the overseas claims. We do not take it as an issue that lacks seriousness. If anything, the complaints we get are that we take too long and resist too many of them. I would not want you to have any sense that we treat them as other than very serious matters, but obviously we do not have the children in front of us. I am afraid that, however well I prepare, for every question you might put to me I do not have-

Q64 Mr Bacon: May I ask you another question about fraud? It does happen to involve overseas issues. Are you aware of the case some years ago, which I brought to the attention of the National Audit Office, involving people coming to this country, getting low-paid jobs, claiming tax credits and then going home again to claim money through cash machines? They were getting several thousand pounds. There were people from different parts of eastern Europe-from Slovakia, certainly, in one case, and I think from other parts of eastern Europe too-and it was on such a scale that it was described to me as industrial-scale fraud. I was familiar with the case because I was briefed by a member of HMRC staff who was sufficiently concerned about it to send me something in a suitably brown envelope. I kept his mobile number on my phone for a bit, until the arrest of Damian Green, at which point I removed it. Are you aware of this case?

Lin Homer: I think that was before my time, but we are continuing to work with colleagues in other agencies on any areas-not just on tax credits but on tax generally-where we believe that there is any amount of serious or organised crime going on. I am aware from both this job and my previous job that that may include circumstances in which there is migrant labour or migrant influx. Yes, we are working in some multi-agency environments. One recommendation in the report is that we should be better at using the data we have to watch for patterns and trends. We agree and accept that, and our use of CONNECT data, which we hear about regularly, will allow us not only to see what we can see with our eyes, but use the machine to spot trends that we cannot see.

Q65 Mr Bacon: One reason why I wanted an answer to the question about the value of the payments to the 7,000 children who live abroad-I share Mr Jackson’s and, I am sure, many other people’s concern about the scope for increased fraud in this area-is to know how worth while it would be for you to employ intelligence officers based overseas in some of these jurisdictions. Do you do that already?

Lin Homer: Well, on the enforcement and compliance side of our business we have investigative skills. We do not tend very much any more to place people in overseas countries. We do when we think there is a need, but they work in networks, so they work very closely with law-enforcement colleagues in the country and more widely, particularly in a European cluster. There are some very live and useful connections in that space.

Q66 Mr Bacon: Do you agree with Sir Richard Mottram that the tax credit system is basically trying to implement the unimplementable?

Lin Homer: I do not often agree completely with what Richard Mottram says word for word but, as I said at the beginning, I think this is an overly complex system.

Q67 Mr Bacon: It is designed in such a way-you more or less said this earlier-that you are bound to have errors on a large scale, and it is bound to take a long time to fix them. The Chair said earlier that she was surprised that the fraud level was so high, and this is one area where I disagree with her because I was not surprised. It is almost bound to be like that because of the design, is it not?

Lin Homer: The design is very complex.

Q68 Mr Bacon: Can you name any other Department where 8.1% of expenditure disappears through fraud and error in this way?

Lin Homer: I can’t, off the top of my head.

Austin Mitchell: What about our agriculture policy?

Q69 Mr Bacon: Well, perhaps. In your capacity as an accounting officer, have you ever asked Ministers to give you a letter of direction before you continued to try to implement this unimplementable policy?

Lin Homer: I have not since I arrived, because I have asked Nick, who is my DG in this area, to give me advice about what we can do to improve the system. As the Chairman said, we have it for only a limited period, and my view is that you have a right to expect me to look hard at what systems I can make. We have made a number of proposals for change to Ministers, and so far they have agreed and accepted our advice. That is where I would start before going marching off asking-

Q70 Chair: I think we should be careful here, Richard, because there is very little fraud; there is error.

Mr Bacon: Indeed. I absolutely accept that some of it may be error, but if we have a system that produces huge errors, the same questions arise.

Lin Homer: Absolutely, so for instance we now have some good checklists on the website. Sue talked about how difficult it can be to work out whether you are in a relationship.

Q71 Mr Bacon: What? Really? Can you explain that?

Lin Homer: Look on the website. We have a disability checklist and there is a very fair challenge that says can we use our own data better, so we are asking the same questions about child credit that we have always asked about child benefit.

Q72 Chair: It is mad that child benefit information is not used to determine the child tax credit.

Lin Homer: I am absolutely determined that we should use our data better, and our front-line staff share our interest in improving this. They become terribly frustrated that, on the one hand, individuals who are entitled to money and need it do not get it, and on the other hand, people-

Chair: Okay. I will go to Austin and then Meg. She talked about not using the child benefit information, and you will do that in future. I know you were not there, but it is madness that it was not used from day one as basic data in what I think we all agree is a ridiculously complex system. It is madness, and I have no doubt that we will be critical of that. I am glad that you are going to change that in future, but it is mad that it has not been done to date.

Q73 Austin Mitchell: I won’t be long. I became a bit sympathetic towards your problem when I read in the Report that you have made little progress in two categories: undeclared partners and work and hours. They are very difficult to deal with, and accounted for losses of more than £1 billion.

Taking undeclared partners first, when I first started as a raw recruit at Yorkshire Television, we got information from the Barnsley Claimants Union that officials from the NAB were spying on single women in the early hours of the morning. We used to go with a film crew to film the NAB officers lurking outside these houses. If a male came out and went to work, they would go in and see whether the bed was warm. If it was, they would assume cohabitation, which was very naughty in south Yorkshire. Unfortunately, when the NAB officers saw us, they all ran away and Yorkshire television refused to pay the overtime bill for all these crews going down to film in Barnsley at 5 o’clock in the morning, so we didn’t succeed in any way. It is impossible to discover undisclosed partners, isn’t it? How do you do it?

Lin Homer: Part of what the CAB was talking about-

Austin Mitchell: Do you do it by people snitching?

Lin Homer: Part of it is that we have run a project in this area over about the past 12 months, which we think has been quite successful. Perhaps I can just ask Nick to outline what we have been doing that we think is making a breakthrough without the escapades you described, Mr Mitchell.

Nick Lodge: It is difficult, I absolutely agree. Since December 2011, we have been trying to address this issue in a slightly different way. We look at our internal data to see whether that suggests anything about another adult possibly living at the address.

Q74 Chair: Do you cross-check with PAYE data?

Nick Lodge: With all the data that we have. We use the CONNECT system, but we also use third-party data. We use credit reference agency data and ask the agency to see whether it has any financial information that may suggest that there is another adult in the house. Where it looks as though there is a significant chance that that is the case, only then do we write to the claimant to ask them whether they should in fact have claimed as a couple rather than as a single person.

Q75 Austin Mitchell: You can’t tell us how many cases you have detected in this way.

Nick Lodge: I can tell you the numbers. From December 2011 to the end of January, we have run about 108,000 of these checks. In approaching 40,000 cases, we have amended the claim from a single person claim to a couple claim, or stopped the single claim. Although your previous witness, Sue, explained the difficulties in a minority of cases, in the large majority of cases we are actually finding that process to be quite straightforward to run through with the claimants. Sometimes the conversations can be quite difficult because, as you described earlier, it is an issue of personal circumstances, but we are finding in quite a high proportion of cases that the information we have from both our own sources and the credit reference agency tells us in practice that somebody else is living at the address.

Q76 Austin Mitchell: I won’t pursue that point any further; I’m not as prurient about it as I used to be when I was at Yorkshire television.

Earning fluctuations is an area where, in my experience of the cases that come to me, it must be your fault. Earnings fluctuate-we have an increasing incidence of part-time working, people are asked to take wage cuts and work fewer hours, and other people work for agencies, with great fluctuations in the amount of work they are allocated. Even if they notify every detail, it is still very difficult to keep up with it. It seems to me that you are so bureaucratic that you often ignore what people are sending in. I cannot see how it is going to be dealt with, because there is so much fluctuation and you are so slow to react that you are going to get errors that are not the fault of the people. It is not fraud; it is just human error. Why can’t you deal with it by having longer periods or wider margins, or whatever, rather than by nit-picking reporting?

Nick Lodge: The tax credit system, as was described earlier, was designed to be a very responsive system. As people’s circumstances change, including their income, their payments can change. Income is based on tax year income, so it is an annual figure. It is part of the complexity that we mentioned earlier that people have to try to estimate their annual income. Quite a good chunk of the overpayments we mentioned earlier are generated by us finding out about changes in income at the end of the year through the tax credit renewal cycle, rather than during the year. So we respond when people report a change, including a change in income during the year, and we respond relatively quickly. It can be very difficult for people to estimate what their annual income is going to be for the reasons that you just described.

Q77 Chair: May I pick up on that? You respond administratively. This was one of the issues that got under my skin a little when I read this report. You have cut your staff by a third because you have had to. If people ring up and report a change of circumstance, that is administratively dealt with. I assume you have cut the very staff who could do even a random check or a more sophisticated risk-based check to see whether the new information on the change of circumstances is correct, or whether or not you will end up overpaying. Am I right? It is just literally an admin job. Somebody rings up, says "I am earning less" and you just shove that into the admin system.

Lin Homer: No, I don’t think that is right.

Q78 Chair: Where are your staff cuts coming from? You have lost a third of staff in this area.

Lin Homer: There is something that Mr Mitchell said that I want to come back to. First of all, what we have done-and I think the NAO recognises this-is to learn from the way we do interventions and use much more targeted and quite interrogative interventions. Both Nick and I have sat alongside some of his tax credit contact centre helpline people who both check the data that is available to them and ask questions in a way that does more than that. But in the main, and certainly until we have RTI, we are waiting for that information, because it is not provided until the event.

Mr Mitchell asked if we could have wider margins. We could, but that costs the country money. An example is a letter I signed this week, explaining to an MP why an individual family owe us tax credit repayments. This is a situation where, when they claimed they declared an income of about £43,000 and when they validated it at year end, they admitted their income had gone up to £75,000. Under the old system, the money they received would still have remained with them. Under the changes to the system, a variation of £5,000 makes the claim subject to change. But it matters that claimants see for themselves that a change of nearly 100% in your income should-

Q79 Chair: That is an extreme case, Lin. May I have an answer to this question? Is it true, because it is my experience as an MP, that when a client rings up and reports that their circumstances have changed-I am not talking about somebody who presumably did not report or did not report the truth-that their report is not properly checked and is administratively handled? Yes or no?

Lin Homer: No, I don’t think that is true.

Q80 Chair: Well, is it or isn’t it true?

Lin Homer: No.

Q81 Chair: Well, if it wasn’t true you would never see any mistakes.

Lin Homer: But we do a number of targeted interventions. We don’t check every change of circumstance that is represented to us.

Q82 Chair: How many do you check? What proportion?

Q83 Fiona Mactaggart: Not only do you check it but you automatically renew a very large number of cases. That is quite clearly, from this report, a source of a lot of fraud now.

Lin Homer: Yes.

Q84 Chair: So how many do you check? What proportion?

Lin Homer: We did a million interventions last year.

Q85 Chair: What proportion do you check?

Lin Homer: About one in six.

Q86 Chair: One in six interventions. The ones you automatically renew, you don’t check. The others that ring up and tell you that they have a change of circumstance, how many of those do you check? What proportion?

Lin Homer: It depends on the risk that they fall into.

Q87 Chair: I accept that it depends on the risk. It is very sensible to have it risk-based, but what proportion?

Lin Homer: We get 6 million changes in circumstance a year. We check. We do about a million interventions. Not all of those lead all the way to a full investigation.

Q88 Chair: No, but perhaps Nick Lodge can answer it, because you have some stats. It is clear in the Report where the problem lies. Out of those people phoning in, my constituents ringing in, telling you that they are earning more, or less-probably more-what proportion, on a risk-based system, do you check?

Nick Lodge: Our contact centre advisers-most of these changes are phoned through to us-will talk to the claimant and do their very best to help make sure that the-

Q89 Chair: I know. That is an admin job.

Nick Lodge: It is a little bit more than an admin job.

Q90 Chair: How many do you check? Give me an answer. If you do not check-

Lin Homer: We disagree with you about what is being done, Chairman. I am sorry you do not like that, but that is the position we are in. At the point when they call, we do not just write a figure down. We have a conversation with them. Those conversations will validate a lot of the cases there and then. You could take that as 100% checked.

Q91 Chair: So how many do you challenge at that stage? If you are saying you check them all, because you check the phone call, what proportion do you challenge?

Nick Lodge: Perhaps it would help if I gave an example.

Q92 Chair: No, not one example. I want a systemic thing.

Nick Lodge: An example of the checks that we carry out. At renewals, which is the point at which people phone up to tell us their income for the previous year, as we go through those calls this year, which will be between April and July, our contact centre advisers will have access to the system that holds pay-as-you-earn data, and they will take some time during that call to check that the information that is being given is consistent with the information that we have had from the employer.

Q93 Chair: So how many do you challenge? What proportion get challenged?

Nick Lodge: They will get challenged at a different point in the process. We can pick them up for a later intervention. We do a scan, a comparison, between the information that we have from employers and the information that we have from claimants, which forms the basis of their awards. We run those checks, and, where there is a discrepancy, we pick those cases up for a check.

Q94 Chair: What proportion? Of the people who give you information, in what proportion is there a discrepancy?

Lin Homer: The interventions carried out-

Q95 Chair: No; please. You can come back in a minute. This is showing to me that you do not have that basic data, whether it is about Europeans or anything. You do not have the data that you need to do sensible things with the money.

Lin Homer: I think we do, but it is trying to give it to you in a way that satisfies you. In 2011-12, we did 1.929 million checks and interventions. Some of those will have been very simple and some will have been very deep. Some of those will have included the 100,000 deep checks we did into undeclared partners, which led to 40,000 being finished. Some of them will have said, "Unless you give us some more information about your child’s education, we are not going to pay you the education element, but we will pay everything else." All of those are valid interventions.

Q96 Chair: Let me just challenge you on that, Lin. Page 19, paragraph 2.10 is really interesting. Interventions increased from 123,000 in 2008-09 to nearly 2 million in 2010-11, so that is a twelvefold increase in interventions, yet the cash in doubled. So you sit there thinking you haven’t a clue, when you are intervening, and your resources are going down.

Lin Homer: I think we would accept that we have learnt quite a lot about which interventions work and which do not.

Q97 Chair: You haven’t learnt. That was 2010-11. We are in 2013. You might just about have the figures now.

Lin Homer: Well, and in the preliminary work that we have done for the subsequent year, we think that 90% of the error prevented comes from 50% of the cases. If we could target as well as that on all the cases, obviously the return, which you rightly said is not as good as some of our other areas-it is still roughly five times return for the money invested-could be made better. Nick and his team are focusing on, if anything, fewer interventions in absolute terms, but more targeting, and that is one of the things that I think we have learnt over the past three or four years.

Q98 Chair: I am going to comment now, then I will go to Meg, Fiona and Stewart. My comment is that I do not think you have the right data. The twelvefold increase in interventions leading to only a twofold increase in money shows that your interventions are not right. I dread to think what the fact that you have a third of the staffing means.

John Thorpe: Just on the data, figure 11 in the Report shows the impact of interventions over the period of the strategy, and that does have 2011-12 data in.

Q99 Chair: Where?

John Thorpe: Figure 11 on page 31. That tracks and it also shows the mix of interventions between those that are prevention at the point at which new claims come into the system, and the balance of effort around what the Department describes as maintaining "cleanse" activities, which are awards that are already in the system where they go back and check that the details they have are correct.

Q100 Chair: I don’t think that that in any way undermines the point I was making. I am sorry to say that.

Lin Homer: Chair, we would accept that there is more we need to do. I tried to start by saying that our view is that we have to get our interventions better, more targeted and more effective. There is no question but that this is an important area for us to focus on. That is why I started by saying that your recommendations are fair and we believe that they are things we should act on.

Chair: I don’t think they are tough enough.

Q101 Meg Hillier: I want to go right back to Mr Lodge’s comment that, of the £1.6 billion that we are talking about, only £40 million, or 2.5%, was the result of error by HMRC. We heard earlier about, and we have all experienced-if not directly, then through constituents-problems with the helpline and advice. Can you quantify what mistakes are being made because of bad or contradictory advice from advisers at HMRC? The evidence we heard was quite worrying.

Nick Lodge: The way that our contact centres operate and the way that they measure whether they are giving the right advice tells us that, for the most part, they are. That is not to say that they do not make mistakes in what can be a complicated area, because mistakes clearly are made. The quantification that I mentioned earlier arises from the sampling that we do. After we have finalised all the tax credits awards, we take a statistically valid sample and we work those cases through. It is from that sample that we derived the £2.3 billion level of error and fraud, and from that same sample that we look to see how the errors are arising. That does not specifically reflect anything to do with the advice that our contact centres might be giving. It is a separate set of figures entirely.

Q102 Meg Hillier: But if that advice is not right, it can cause the error in the first place. Even if it is a small percentage, the impact on the individual and their family can be huge. I am sure that that is not lost on you. How do you make sure that where you have evidence-you have got quite a lot of evidence from the CAB, which I hope that you will look at in more detail and act on-that does not recur, and that you have robust measures in place to retrain staff, or to support claimants who have had this advice in error?

Nick Lodge: We have very regular meetings with the CAB and with Sue and her colleagues, and they are a great help to us in alerting us to issues that come across their desks, some of which Sue mentioned in her evidence.

Q103 Meg Hillier: So you don’t mystery-shop yourself; you leave it to organisations such as the CAB?

Nick Lodge: We carry out a pretty rigorous quality check, picking up quite a large, statistically valid sample of calls; all our calls are recorded. We check those against the guidance that our contact centre advisers are supposed to follow to ensure that we are giving the right advice. The figure is not 100%, but it is in the high 90s. Where we find that we are making mistakes, that gets fed back to advisers very regularly. There are daily sessions, we update the guidance and so on. That is quite a rigorous process. That is not to say that we do not make mistakes, because we do, but we do our absolute utmost to minimise those. We understand the impact on people when we make mistakes.

Q104 Meg Hillier: Thank you for that. I want to move on to the wider point about the Government programme "Tell us once", which Lin Homer might want to comment on. The idea of that programme is that somebody will ring up and tell one bit of Government about a personal circumstance, and that will be enough to tell all of Government. We still hear, however, that your Department does not make the match between child benefit and child tax credits, which seems crazy, as the Chair highlighted earlier. As a non-ministerial Department, do you lobby DWP and the Cabinet Office about "Tell us once"? It seems to me that you are one Department where it would make a very big difference to your day-to-day operations.

Lin Homer: We do work very closely with DWP, don’t we?

Lin Homer: We do work very closely with DWP, don’t we?

Q105 Meg Hillier: Can you just make a comment about where we are with "Tell us once", because it seems to have gone into the sand a bit? The Cabinet Office vaguely told us on Monday that they were looking at certain areas of it again. DWP say that it is done and finished; because they are dealing with universal credit, you get the impression that it has fallen right down to the bottom of their agenda. Whichever Department leads on it, it has a bigger impact on many other Departments, like you. You work very closely with them, but are they actually going to deliver and help HMRC? On what time scale?

Nick Lodge: We do share information with DWP. I think there is more we could do around the "Tell us once" agenda; I am sure there is. However, for example, we share data on disability, which is one of things that the NAO Report highlights, so we share data in that way. I should just say that I absolutely agree about the child benefit data, but we now regularly compare the child benefit system and the tax credit system.

Q106 Meg Hillier: Is that automated now?

Nick Lodge: It is not automated. The child benefit IT system is, I am afraid, extremely old-well over 30 years old-and there is no automatic link between the two systems. We do regular scans that compare the data in the two systems; it is not ideal or perfect, but we do make that check.

Q107 Meg Hillier: Okay, so "Tell us once" is not there, and neither is the technology; that is depressing all round really. I want to touch on the issue of the overestimate of the impact of your interventions on the level of error and fraud. You revised your estimate down from £1.4 billion to this £480 million in paragraph 9 of the Report. The NAO conclude that part of the reason was that you were not really testing your assumptions by looking again at cases after you checked them. I wondered why you were not doing that, or if you agree with that.

Lin Homer: Just as an overview, I think the point that we alluded to earlier-that we have this long lag before we can absolutely check the data-does mean that the methodology that was started was tested and based on experience in other tax systems and, essentially, it had components; one chunk was a direct impact, one chunk we expected to be a deterrent impact on others, and another chunk we expected to be a kind of continuing impact on that claimant. Both the latter two produced far less than we expected. The direct impact happened as we expected, the deterrent we found much harder to show, and the longevity of the impact on the accuracy of that claim has also been shorter than we expected.

Nick has been doing a lot more work on the interventions we have now, and the ones we have planned for the future, to try to ensure that we are testing those. One of the things that I have been talking about with the NAO is whether we can, in a sense, do more open-book accounting for these interventions as we go along, so that our NAO team can probe and test those as we go. We are testing them validly ourselves, but there is a question for us about whether we can get more real-time sharing out of that. It also means that we think that we should test and pilot some of these interventions and then do deep evaluation-again, I think that is backed up in one of the recommendations-so that we are not waiting so long to look back and make a judgment. That is work that Nick has been doing since he started last year.

Q108 Meg Hillier: So you are going to be doing this real-time tracking?

Lin Homer: We are beginning to do much more of that.

Nick Lodge: We are beginning to do it. We have already selected about 500 cases to follow through from our renewals process this year. We will select more than that, because we agree with the recommendation from the NAO that we should.

Q109 Meg Hillier: Do the claimants know that you are following them through in real-time?

Nick Lodge: They will know if we intervene in a particular case, because we will need to check with them to see what has happened. Part of what we have learned here from the analysis is that awards become incorrect again far quicker than we thought they would. Therefore, if it is errors, people are making the same mistakes again, despite us having intervened earlier. Part of the learning for us from that is the need to understand precisely why that happens. We have commissioned an independent company to do some social and economic research, to talk in depth to some claimants to establish why that it is, and to learn what we can do differently to try to prevent that happening again. I think the key piece of information that has come out of this for us is the need to intervene more frequently to help claimants keep their awards correct.

Q110 Meg Hillier: Can I return to my question about whether they know you are doing it in real time? Psychologically, if you know that HMRC is watching, over your shoulder, every step you take, it is a bit like having a teacher over your shoulder: you will be a little more careful, naturally, about checking than if you think it will be kicked into the long grass and that you will not have your next intervention with HMRC until next year. If I were claiming child tax credits now and you were tracking my case, would I have any idea of that direct tracking, or would you just intervene at a point where you thought you needed to?

Lin Homer: We have used behavioural insight in this area. We could share with the Committee some examples of letters that we use. We tone the letters to signal to people that we are getting more interested in them, if I can put it that way.

Q111 Chair: The answer is that you do not tell the client.

Lin Homer: We do, actually.

Q112 Chair: Or you only tell the client if something is wrong.

Lin Homer: We do not say to all 6 million, "We’re watching you every step of the way," because it would not be true and we should not be generating that level of anxiety in everybody all the time, but when we write, we try to signal much more directly. We might write a letter that says, "We have another source of information that suggests there is something wrong with your claim. Would you like to check it for an error and tell us what it is?" In another one, where someone has made repetitive errors, we might go further and say, "This is the third time you have made an error. You need to remember there is the risk of penalties if you keep doing this."

We are trying to use much more targeted letters. This is the general behavioural insight that we have talked to you about. Our experience is that if the claimant has a sense that we are close to them, it activates more reaction from them. So we have a range of letters, not just confettied out but aimed at the circumstance that presents itself. That is one thing that we have learned from targeting.

Q113 Meg Hillier: What are the time scales? It is always good to pin people down. Given that you agree with the NAO’s recommendation, when will we see a step change in what you are doing? You are talking about a few hundred at the moment.

Lin Homer: We will select a few hundred, follow those through, and learn from them what causes an award to go wrong. We should be getting that information through this coming autumn.

Q114Chair: Just for the record, you are committed to £8 billion in savings by 2015. That is what you committed to.

Fiona Mactaggart: I think the killer paragraph in this Report is paragraph 19, which says quite specifically that you will not achieve the target to deliver £8 billion. How much are you going to deliver?

Lin Homer: Well, what we have confirmed is that we did not hit the target.

Q115Chair: How much?

Lin Homer: If we stayed where we are now, on the current estimates, we are probably only on target for a third, but if we do some of the things that we are trialling at the moment and move on to believing we can industrialise those, that figure may go up. What I have asked Nick to do is make those judgments as he undertakes each of these interventions and changes what he is doing. That is why I am saying we think we have to do more. If we stay where we are, it will be somewhere between £2.5 billion and £3 million, and clearly the Chancellor would like us to get much closer to £8 billion.

Q116 Fiona Mactaggart: So before this system ends, you are due to pay out some £100 billion.

Lin Homer: We have paid £30 billion a year.

Q117 Fiona Mactaggart: Up to 2014-15. And you cannot yet tell us how much more than, say, £2 billion to £3 billion you are going to save from fraud and error.

Lin Homer: I think experience shows that putting an estimate on this without the evidence to base it on is not a very sensible thing for me to ask Nick to do.

Q118 Fiona Mactaggart: I think figure 8 in the Report makes that very clear, doesn’t it?

Lin Homer: Yes, so what we are aiming to do is clearly set ourselves the target of getting closer to the original estimate if we can, but do so on the basis of tested and evaluated proposals that are worth investing in, rather than a confetti approach. As I said earlier, what we are suggesting is that we account for those steps as we make them and invite, within our regular relationship with the NAO, their comments on that, which, if generated, would, I have no doubt, cause another hearing with you, if there were a point when the NAO felt that we were being dilatory or not trying hard enough. Obviously, you can rest assured that if I think he has a cracking idea, I will come back and tell you myself.

Q119 Fiona Mactaggart: I want to go to figure 9, which looks at the areas of error and fraud, what proportion, and what has happened in them, if you like. We know that the 2008-09 figure might be suspect to some degree, but let us take these figures as correct. There are three areas where there has been a significant-over 2%-increase. There is a 14% and, in one case, a 96% increase in error and fraud. The 96% increase is in relation to children. Reading the Report, it seems to suggest that that is overwhelmingly automated repeat claims. Am I right, or is there some other cause for that very significant increase in the amount of error and fraud in relation to children cases?

Lin Homer: I think it is shown quite visually in figure 1, as well, where the relative level of risk is shown from 2008-09 to 2010-11. As you say, cases involving children have gone up. Nick, do you want to say what we think is behind that?

Nick Lodge: Some of it, as you rightly say, is due to automatically renewing some claimants through that annual renewals process-not all of it; some of it was simply because we did not carry out enough checks in that area for that particular year. That is one of the areas where we have moved to increase the number of checks that we are going to carry out, and where, for tax credits in future, we will mirror the kind of approach that we take in child benefit, where we ask parents on an annual basis to reconfirm that young people are in continuing further education. That is one of the areas where people simply largely forget to tell us.

Earlier, Sue talked about the different approach with child benefit. We are going to mirror that approach, continue to carry out matches of the child benefit and tax credit systems, carry out further interventions, and talk to the Department with responsibility for employment about whether we can do some checks via their data as well.

Q120 Fiona Mactaggart: How much do you expect that action to save?

Nick Lodge: We are expecting to see a continuing downward trend overall.

Q121 Fiona Mactaggart: No, at the moment it is an upward trend. It is not a continuing downward trend. That proportion has been an upward trend.

Nick Lodge: It is, and we will get the latest figures from the sample in about May. We will then be able to see the effect of those interventions. What we are seeing day to day, as we do the checking, is that in about 35% of the cases in which we are checking whether there has been a change of circumstances, there has been a change, and we have been able to correct that. I do not have a quantified figure for you, but we will have that in May or June when we get the sample result.

Q122 Fiona Mactaggart: Have you given a quantified figure to the Chancellor for how much you are going to save?

Lin Homer: He has not given a quantified figure to me yet, because I have taken the view that I would rather have something I could place some reliance on than a guess.

Q123 Fiona Mactaggart: That looks to me, from this Report, to be kind of necessary, doesn’t it? There are these three top groups. You have talked a bit about what you have done in relation to the undeclared partner. Then there are children, and work and hours. I suspect that the work and hours errors are largely to do with the radical change that is going on in the workplace at the moment.

Lin Homer: Yes, and we think RTI will help.

Q124 Fiona Mactaggart: That is to do with people having unpredictable hours and so on. That one you are probably not going to be able to change hugely, because people are on zero-hours contracts; they do not know how much they are working. I think that figure will stay high. It would be shocking if the other two, particularly the one relating to children, is at anything like the proportion it is currently at. It seems to me that you created the problem with the automatic renewal system.

Lin Homer: I think we share your view that the top three are where we need to put effort. We think RTI will help us with work and hours.

Q125 Chair: RTI is down the road, Lin. What are you going to do between now and 2018-19?

Lin Homer: We think RTI will help us-not DWP-with hours in 2014.

Q126 Chair: You had a quarter wrong in your pilot.

Lin Homer: No.

Q127 Chair: According to PQ answer, a quarter of the cases that were handled in the pilot included an error.

Lin Homer: My point was that I do not agree with Fiona.

Q128 Chair: You don’t depend on RTI; you are miles from there.

Lin Homer: We are not just going to ignore work and hours; that is my point. On children and undeclared partners, we would accept that that is where we need to focus. As Nick has said, the programme of work that we have been doing on undeclared partners over the past 12 months has led to some quite significant results. The question is, can we industrialise those-can we move from 100,000 cases to similar checks on many more?

Q129 Fiona Mactaggart: But if you industrialise them, you have to have a more effective and faster appeals system. We learn that people are living on air while they are waiting for the result of an appeal in cases where there have been errors by you.

Lin Homer: That is why I think it is important that we test and validate these approaches before we industrialise them, so I am in the same place as you.

Q130 Mr Jackson: A slightly gentler question, before you get your gin and tonic: have you looked at any comparator systems across the world that are analogous to our particular system? What lessons have you learned from international experiences, in terms of efforts to reduce fraud and error in a similar system? I know that this is a unique system, but there will be commonalities across different countries. Have you had an opportunity to look at that to try to improve the system?

Lin Homer: We are trying to do some benchmarking generally-I have talked to you before about that-but I am not sure whether we have done any in this specific area.

Nick Lodge: We may have done, quite some years ago. I would need to check to see if we have done anything more recently.

Q131 Chair: Who else has a system like this?

Lin Homer: Well, it is quite unique.

Nick Lodge: There are some broadly similar systems in the States and Canada, but they are different because they are tied in more with the tax system and universal filing, whereas this is different in that claimants have to make a claim. So I would say that it is fairly unique.

Lin Homer: But we do believe that there is always value in looking at other people’s systems. Edward Troup, the second permanent secretary, is looking generally at how we can more regularly and more transparently benchmark ourselves. We will ask him that question-whether we can find valid systems that we could compare ourselves to, perhaps a bit closer to home.

Q132 Chris Heaton-Harris: I have what I hope are really soft questions.

Lin Homer: You just say that to lull me into a false sense of security.

Chris Heaton-Harris: No, I actually feel very sorry for your staff, who are having to deal with this system; although it is as fit for purpose as it is going to be, you would not start from here if you were starting again. I know you have some really hard-working staff behind the scenes who are trying to make this work.

Mr Jackson: That is soft enough, Chris.

Chris Heaton-Harris: I am just being nice. I apologise for missing the first session; both Mr Bacon and I were giving evidence to another Select Committee. That had a very odd feel to it.

Lin Homer: You are feeling generally sympathetic to me today, then.

Q133 Chris Heaton-Harris: Yes. You have got me on a good day. I spend a lot of time in a citizens advice bureau in my constituency-I run my surgery from there-and a lot of my basic work as a Member of Parliament is trying to help people work their way through your system and fix their tax credits, when an error has come about and you have written to them. I assume you saw a copy of Sue’s flow chart.

Lin Homer: Yes. We talked to Sue while we were waiting.

Q134 Chris Heaton-Harris: Mr Lodge, you talked about the meetings that you had at the highest level. I am wondering whether any of that filters down to your staff, who would appreciate the advice and help that citizens advice bureaux give-they are doing part of the work of your service. Can you give me an example of that filtering down? From my dealings, I am not convinced that it is trickling down the system as well as it should.

Lin Homer: I will let Nick give you a bit of the detail. I want to echo your point. I think our staff try to keep the balance in this system; they really do work very hard, both to remember that they are dealing with families and human beings and to work as well as they can, and to remember that they have a big responsibility for making the system work. They work extremely hard, and probably give us more grief than you do about trying to get the system right. We are trying to ensure that they feel supported in bringing up their ideas so that they don’t have to send them to Mr Bacon in a brown envelope.

We are working increasingly well with the voluntary sector, I think, although I am sure there is always more that we can do. While we were waiting to come in, we were talking with Sue. She has visited our front line before, and we have invited her to come again, to listen into some of our calls and to give us advice on what she thinks we could change. We also suggested that she might like to do a session with our front-line staff to talk about what her staff in various CABs see.

In a wider area, we have been working more generally with tax agents and the voluntary sector, and sending our staff to sit in on surgeries with agents and with voluntary sector groups. We then do the reverse, and bring them back in. We have a joint programme where we are working together, and we try to pick areas and work on those. Most recently, we have been looking at debt management. There has been a big piece of work involving both the voluntary sector and agents working with our debt agents, watching what we do and commenting on it. Our debt people then join them. That is a recent example, which we are just about to get a report on. So we get the value of that. We also have an escalation process involving Liverpool, haven’t we?

Nick Lodge: Yes. We work very closely with CABs, as I mentioned earlier. Above and beyond the national meetings you mentioned, the CAB has set up Liverpool as a local office with national expertise in some of the tax credits matters that we have been discussing. We regularly meet people from the Liverpool citizens advice bureau at a local management level. That is an escalation route for any particularly difficult cases that come to their attention, but also for a general discussion about how the process is working and improvements that we might make.

Lin Homer: That is escalation for all CABs, not just for Liverpool.

Q135 Chris Heaton-Harris: I am glad to hear that, because one of the questions I wanted to ask was about that sort of interaction. In my local CAB in Daventry there is a fantastic lady, Anne Bamfield, who has been doing that work for decades, I think-it is a very long time. She has a wealth of expertise. She has seen the system from its birth-from its conception, all the way through. She has seen all its wrinkles and problems, and she is probably more frustrated about it now than she has ever been. I wonder about the interaction, the swapping of knowledge and the personal contact that those sorts of meetings could encourage. She and people like her from citizens advice bureaux, instead of having to sit with thousands of clients of the CAB, would probably be better spending an afternoon sitting down with your staff, who also want to do the best thing for your clients-I understand that. You can identify issues and points that you do not understand when people phone up. It is even worse when the phones don’t work, which happens a bit too regularly-I saw that point somewhere on here-because people are pretty desperate at the point that they pick up the phone to you. I think you are missing out on a chance to use that expertise and empathy, and build up those personal relationships.

Lin Homer: Yes. We absolutely think there is much more benefit from us working more closely. We will make sure that lady is one of the names we talk to Sue about. I do not think that anything other than good will come out of our getting those local discussions going. We need to work out when there is a policy change, an admin change or a resource issue, so, instead of solving one case, we solve many with those kinds of intervention. That is when it becomes really powerful.

Q136 Chris Heaton-Harris: I have one final question. In the list of problems on this sheet, would you be so kind when you are writing the notes to send us a note on how you have dealt with those sorts of problems. It can be generic, but I would be much more comfortable knowing that you have not just listened to us, but someone has had to sit and think, "We have dealt with that problem. The third one is phone numbers to compliance not working for a long period. Yes, we have identified that; this is the solution that is being put in." That problem especially happens way too regularly for anybody’s good.

Lin Homer: Very happy to. We will keep the same format and add either what we have done or what we plan to do, and distinguish between those two.

Q137 Mr Bacon: I have an endorsement and a question. The endorsement is for talking to the CAB. I was astonished-I am not surprised by anything any more after a few years on this Committee-when I first discovered that my local CAB was receiving calls from my local DWP to check what the rules were. There is a wealth of knowledge, and I am glad you are focused on that. I would perhaps encourage you to do more of the local stuff as well as the national stuff. As you say, the trick is then, having got the local problem, to nationalise it. It is at the very rock face that you are most likely to encounter the real nature of the problems. I welcome that. I am very glad you are doing it. Well done.

My question is about your comment on behavioural insights. You said in passing, "We use behavioural insights in writing these letters". You were specifically talking about potential fraud and error in relation to tax credits. Gus O’Donnell has talked about this to us in the past when he was still in office, although he was talking about that in relation to HMRC more generally. You write to taxpayers in Dorset and remind them how many of their citizens and neighbours have already paid their taxes and so on. Do you use behavioural insights in the management of HMRC and your own staff?

Lin Homer: Yes, we do, but I have recently asked myself the question whether we use it as well internally as externally. You have challenged us on how low our engagement scores are, so we have been trying to respond to what we know is causing those behaviours in the way that we interact. For instance, that is leading us to try to problem-solve with groups of staff throughout the organisation. In much the same way as Mr Heaton-Harris talked about doing joint work with the CAB, if we have an area where staff are saying that something is not working, we try to draw in and involve them.

Q138 Mr Bacon: You mean that you didn’t before.

Lin Homer: Not enough, I think.

Q139 Chair: Can I intervene on this one? Page 9, paragraph 18 says: "HMRC has not yet developed a detailed picture of how claimants behave". The report that you have signed off, Lin, suggests that you have not done what Richard is suggesting would be sensible.

Lin Homer: Well, not yet developed a detailed picture-

Q140 Mr Bacon: What I am really asking is, "Have you developed a detailed picture of how HMRC staff behave when they feel completely frustrated because no one is listening to them?" Wouldn’t behavioural insights be very powerful for improving that, as well?

Lin Homer: The answer is yes to that, and the answer to you, Chair, is absolutely. We have not done enough. I am saying that we have begun to use behavioural insights with claimants, and we are beginning to do so with staff. One of the things it does teach you quite quickly is that you do need your problem-solving to involve the right people at the beginning, then you need your evaluation to involve them. We will relatively soon be producing our first annual report, as a result of our tax assurance changes. You will recall that one of your criticisms on the big tax disputes case was that the experts involved early on were not involved all the way through. We are now retaining the input of those early experts all the way through to the final commissioner decisions.

Q141 Mr Bacon: This is all good news. It is music to my ears. It has plainly been very prevalent, and I am interested in what possible world view could have developed that said-as apparently from your last sentence is the case-you didn’t need to have the right people involved from the beginning? How did you get to that place?

Lin Homer: I am not suggesting that we didn’t all of the time, but what is plain is that, with claimants-

Q142 Mr Bacon: I am not just talking about you. It is a common problem throughout Government Departments. I think you did more or less suggest it, actually.

Lin Homer: I think that we need to do much more of it. What our internal review of our engagement work said is that we were sometimes-not always; in an organisation as big as ours there are always excellent leaders and less good leaders-more directive and less inclusive than we needed to be. We are trying to learn that lesson. I do not think you can do that as perfectly as you want in a big organisation, but it is still possible in a big organisation.

A recent review of one of our systems that our staff don’t like at all has involved more than 1,000 of them in commenting on and putting views forward. At the end of the day, it will be a smaller group than 1,000 that has to decide what you do with those outcomes. They come down to me, but you can then explain very carefully to the people who have been involved what decisions you have made and why. That approach hasn’t been absent, but we believe by doing more of it that we could be a better organisation.

We believe that, by involving the users of our systems and the third party organisations that work with us, whether they are tax agents or the voluntary sector, we can get to solutions quicker.

Q143 Mr Bacon: Do you do 360° appraisal?

Lin Homer: Only for more senior people, but one of the things that I have been talking to the staff about in my staff visits is whether we should do more of that. We have monthly dial-ins when we just take questions. I have one member of staff who does a 360° appraisal of me live on air once a year, so that others can comment. I got five out of 10 from him last year. I am hoping to do somewhat better this year.

Q144 Chair: I am going to sweep up. Organised crime is mentioned. How much is at risk from that? Do we know?

Lin Homer: It is not covered in this report.

Q145 Chair: It is mentioned on page 14.

Lin Homer: The reference is that it is not part of these credits.

Q146 Chair: Oh, I see.

Lin Homer: We are obviously using both our special investigation and our criminal investigation people. We do undertake serious and organised crime. At the risk of annoying Mr Bacon, I do not have the figures. But what we think that we lose to serious and organised crime off the top of my head-

Q147 Chair: Will you let us know in your note?

The other thing was a bureaucratic thing. I know that HMRC is complicated. I know that this particular tax credit is a nightmare in its complication, but when I read that you had 40 different interventions, I thought that your poor members of staff will take 10 hours to get through the manual to find which intervention is the relevant one to deal with my constituent, who is just fed up because she has been asked for another £3,000.

Lin Homer: Do you want to explain, Nick?

Nick Lodge: I think that we actually have more than 40 now.

Q148 Chair: I think that that is a nightmare. You are creating a bureaucracy. The truth is that they are mainly lone parents. They are mainly in work. They may have a different number of children, all that sort of stuff. We see them in our surgeries. They are not that different. If you make it so complicated, you will never get to grips-even with what is a nightmare complex system.

Nick Lodge: You are absolutely right. The main body of interventions we deal with are straightforward ones that you would identify. Writing to people to check the evidence, writing to people to check their current circumstances and so on form the mainstay of the checking that we do.

Lin Homer: To be clear, some of our systems work out which interventions to apply, so we don’t leave the call handler going, "Oh, what shall I do from this list of 40-plus?" The system says, "This is a person who has made an error in this area before. Use this style". We are using our data analytical capacity to guide both claimants and staff where to focus their limited attention.

John Thorpe: There is a national assessment of risk in the tax credits office. They use data to identify groups of claims, which have a particular characteristic and a particular risk. Then they use teams that are dedicated to working to a standard operating procedure to work through those claims, to address that risk. They are not having to make a choice between 40 different interventions when the phone rings. This is managed from the tax credits office.

Chair: I was going to refer you to figure 7 on page 21. It is the IT system. Who provided it? How much did you spend on this IT system that didn’t work?

Nick Lodge: This is the work front risks-

Chair: How much did you spend? Who provided it?

Nick Lodge: It costs something just under £1 million a year to run, and I think it cost about a £1 million to develop.

Q149 Chair: Who was the provider?

Nick Lodge: It is a Fujitsu system.

Q150 Chair: Oh, famous. It didn’t provide what you thought it was going to. You are supposed to be able to verify automatically against other HMRC systems-something we have been talking about-but you end up doing manual checks.

Nick Lodge: This is a system that we run against all new claims before they actually get put into our system. They come in; they get scanned, and then they run through this system. I would argue that it has actually been one of the success stories in trying to stop error and fraud getting into the system in the first place. The system can draw on a raft of data that we hold internally and also externally. It can spot patterns. It throws out for us to check particular types of claim, before we even put the claim into the system so that we can then check with the claimant.

Q151 Chair: So why do you have to have a manual check?

Lin Homer: This is part of the lovely new world of data analytics. The system is saying, "Look at these cases". The million pounds we spent on developing it has already saved us £256 million in losses prevented. That is a pretty good return. It was worth spending. As Nick said, it identifies repetitions and patterns. It identifies bank accounts, telephone numbers or addresses that appear more frequently. We then hand over to a skilled investigator-a human being-who then follows it up. It is an intelligent system to use alongside our people skills, which is what it was designed to do.

Q152 Chair: Does it give you everything you wanted it to give you?

Lin Homer: I am sure-I say this about all of our IT systems-it has been well worth it. I am sure that there will be a son of FEAST in due course, and it will do more and probably do it faster, but it is teaching us about how much we can find out by this recognition of patterns and unusual circumstances. It is helping our people focus their attention.

Q153 Chair: Two more areas, one of which is universal credit. I have information that one quarter of the pilot cases with the IT system for RTI failed. Are you telling me that I am wrong about that? Would you like to write to me?

Lin Homer: It is not a figure that I recognise, so if I could get the detail of the PQ-

Q154 Chair: It is not a PQ. I think that it came from something else. I am sorry. It is one of the pilots you have done to see whether RTI is working.

Lin Homer: By the end of March, we will have had six million employee details in. My belief is that we have had well in the 90s in terms of accuracy.

Q155 Chair: Okay, well write to me on that. I cannot remember. I probably picked it up in the press, if I didn’t pick it up from a PQ.

I feel that you are seeing RTI as a bit of a panacea to deal with a lot of the complexities and difficulties you have. Let me just take one example. Fiona talked about the undeclared partner issue. How will RTI help you in any way to deal with the problems of people not declaring their partners?

Lin Homer: The reference I made to the categories that were more unresolved for us was to work and hours. That area is most beneficial to us, because it will give us regular and accurate information about the variable income of one or both partners.

Q156 Chair: Undeclared partners.

Lin Homer: As we described earlier, we are increasingly comparing data.

Q157 Chair: This won’t help at all.

Lin Homer: It will allow us to track back. Again, some information is coming through one system. One of the things that RTI will do is require employers to have accurate information about the name, address and date of birth of their employee. If some of the information they are giving to their employer is different from the information that they are giving to us, it will come through. That again is third party data, and third party data has been very valuable to us on undeclared partners, but its major immediate benefit will be in working hours.

Q158 Chair: My final question. You told Fiona that the £8 billion that you put into the CSR was the money you had saved from error and fraud. I would love you to report it separately because I think that it would deal with a lot of the myths around this. You will get £2 billion to £3 billion of that at the moment, may be with a bit of luck a bit more. Let us say, with luck, you get to £4 billion. Where you are going to get the other £4 billion that you are required to have to meet your CSR financial targets?

Lin Homer: With no further change, that is the outcome. I do not think that it will be luck. It will be hard work from Nick and his team.

Q159 Chair: Hard work leading to luck, sending your figure up.

Lin Homer: It is early days.

Q160 Chair: Don’t let’s quarrel about my language. You are not going to get to £8 billion. Let us assume that you get to £4 billion, so you are £4 billion light on your commitment in the CSR.

Lin Homer: Our general progression on debt, which we have talked to you about elsewhere, is progressing reasonably well. We are likely to continue talking to Ministers about further work on fraud and error in relation to debt. On the overall position of our commitments around debt, we remain committed to trying to extend and improve.

Q161 Chair: I do not doubt that. You may say you have no plan B, but in all the figures that we get and that Parliament gets, which the Chancellor is working on, he will have assumed the £8 billion you promised. Even with the hardest work and huge innovation, given where you are now, you are unlikely, by 2015, which is not that far down the line, to get to, let us say, £4 billion. So you are £4 billion light, on best predictions.

Lin Homer: We have not yet, but, overall, we have been overperforming in some areas of our CSR targets, including our yield targets. We will keep working, and I suspect the Government will keep publishing their views every financial episode. At the autumn statement, we increased by £2 billion our expected yield around our overall compliance work, so this is an ongoing negotiation with Ministers.

Q162 Chair: The reason why I raise this is that it is so important. This area goes across Government-across everybody. The more you do here, the less we cut elsewhere.

Lin Homer: Absolutely. That is why I want to emphasise that Nick has some projects. We have a new project we will be piloting in the next four months or so, utilising a private sector approach. As I said earlier, if we undertake any interventions that we think do work and can produce an improved process, our plan would be to evaluate and implement them quickly. So we are certainly not calling time on these outcomes yet. As I said, I would prefer to make proposals based on evidence pilots, rather than guesses, but we certainly believe we have to have a really good go at getting as much of this debt in as we can, to make the system high integrity for the people who pay for it and to do our share towards-

Chair: And of course the remittance.

Good. We will no doubt come back when we do the report on accounts next year. Thank you very much. See you tomorrow morning, bright and early.

Prepared 11th March 2013