Public Accounts Committee - Minutes of EvidenceHC 1531

Back to Report

Oral Evidence

Taken before the Committee of Public Accounts

on Monday 17 October 2011

Members present:

Margaret Hodge (Chair)

Mr Richard Bacon

Stephen Barclay

Stella Creasy

Matthew Hancock

Mrs Anne McGuire

Nick Smith

Ian Swales

James Wharton


Amyas Morse, Comptroller and Auditor General, John Thorpe, Director, National Audit Office, Jane Wheeler, Director, National Audit Office and Marius Gallaher, Alternate Treasury Officer of Accounts, HM Treasury were in attendance.


HMRC Tax Disputes

Examination of Witnesses

Witnesses: Stephen Banyard CBE, Acting Director General Personal Tax, HM Revenue and Customs, Dave Hartnett CB, Permanent Secretary for Tax, HM Revenue and Customs, and John Keelty, Director, Finance Planning and Performance, HM Revenue and Customs, gave evidence.

Q230 Chair: Welcome again, everybody. We have a lot to get through because we want to focus on the parts of the Report that we weren’t able to cover last week and we want to deal with the other Report from the NAO, on cost reduction in HMRC. It is appropriate to say in the public arena that we aren’t going to let matters rest from last week. The Cabinet Secretary has asked to see me and I intend to attend that meeting with Richard Bacon from the Conservatives and, I hope, Ian Swales from the Lib Dems. From there, we will decide what further action we’re going to take. We would like Amyas to do some further work, and perhaps, Amyas, you’ll describe that further work. There are one or two questions and then we’ll move to that. In the first instance, would you like to describe the work?

Amyas Morse: The proposal is that we would put ourselves in a position to examine the reasonableness of some of the larger settlements, probably with the benefit of some tax advice, and we would aim to report on that in a confidential-private-session at some point in future. I know HMRC and Gus O’Donnell would co-operate, but it would be helpful to have that confidential session. That might help to satisfy a lot of the issues that are still there.

Dave Hartnett: Chair, if I may, you asked me a question on Wednesday, which I was not able to answer because Mr Mitchell asked a question at the same time, and I fear I turned to his question rather than yours. The question was whether we would talk to you in confidence. Had I answered, I would have talked of the sort of confidential discussion just described by Amyas Morse. We are ready for that sort of discussion.

Chair: Good.

Dave Hartnett: And we will support the NAO in its work.

Chair: Thank you very much. There are just few questions arising out of the transcript from last week that Stephen, and perhaps Richard, would like to put to you. Then we will move on to discussions of PAYE and other matters.

Q231 Stephen Barclay: Starting with the legal advice, which you offered to send. We have not received that ahead of today’s hearing. Could you clarify when that will be received?

Dave Hartnett: The difficulty, Mr Barclay, is that the legal advice is subject to legal, professional privilege. The Government do not disclose their legal advice publicly, save in very rare circumstances. We have said that we would write setting out the legal position. We have been asked to advance the time by which we do that to Wednesday lunch time, and we will meet that timetable.

Q232 Stephen Barclay: Can you advise the Committee on which areas you sought legal advice?

Dave Hartnett: I’m thinking of the legal advice all the way back to 2009. We were going to cover the whole piece for you. Basically, it is legal advice on taxpayer confidentiality.

Q233 Stephen Barclay: Coming on to one of the issues we were trying to get at around the point that you are arguing on individual taxpayer confidentiality, Vodafone, in its own regulatory announcement on 23 July 2010, said that the settlement comprising the £800 million in the current financial year would be paid over five years. In your evidence to us last week, you said that time to pay would be granted only if a company did not have the assets to settle its bill. Given the strength of the Vodafone financial position, why was it given five years to settle its bill?

Dave Hartnett: I fear, Mr Barclay, we are back where I was last week. I cannot talk about Vodafone, but I am sure we should be able to pick up something like that in confidence in the context Mr Morse and the Chair have described.

Q234 Stephen Barclay: This is an area that the company itself has put in the public domain. It made a statement saying it had five years. You have guidelines-also in the public domain-that say time to pay cannot be granted, except if the company does not have the assets. Vodafone’s financial position is in the public domain. I can happily quote its position on its operating profit increasing from £11.8 billion to £12.2 billion, the dividends it was paying and other points regarding its financial strengths. That is all in the public domain. It seems there is a clear breach of your own guidelines, yet that is information Vodafone has shared with its investors.

Dave Hartnett: I am sorry but, even where information has been put in the public domain, I cannot discuss a taxpayer’s position. We will try to do so in the "in confidence" discussions.

Q235 Stephen Barclay: Is it just a coincidence that the settlement with Vodafone was reached the day before the quarterly announcement by Vodafone?

Dave Hartnett: I have no idea. I am sorry.

Q236 Stephen Barclay: So, the fact that that settlement was reached with Vodafone, without written legal advice, the day before the company announced its dividend position and its quarterly returns-that is just a coincidence, is it?

Dave Hartnett: I cannot comment on that.

Q237 Mr Bacon: Mr Hartnett, I want to ask you about the mistake that you referred to last week. That was what you described as "an issue in relation to which a mistake was made, because we saw an impediment to dealing with an issue. In fact, the impediment had been removed. A number of us at the meeting did not know that." You went on to say, "It was actually me who took the mistake to Mr Inglese." When did you first know that there had been a mistake?

Dave Hartnett: Probably four or five days after the meeting.

Q238 Mr Bacon: Just to remind us, what was the date of the settlement meeting?

Dave Hartnett: I think 19 November.

Q239 Mr Bacon: So you were abroad when you found out about the mistake?

Dave Hartnett: There was-I need to check. I may need to correct when I went abroad. I may have gone abroad a week later.

Q240 Mr Bacon: You were in London on the 22nd. On the 23rd and the 24th you were in Liechtenstein. On the 25th and the 26th you were in Switzerland. That was the whole of that week. The following Monday, the 29th, you were in India, and you did not get back until 5 December. In fact, you said to us last week that you went straight abroad after the meeting.

Dave Hartnett: I thought I had done, Mr Bacon.

Q241 Mr Bacon: What I am trying to figure out is-because you say that you identified the mistake-when did you first know that there had been a mistake?

Dave Hartnett: I believe-I will have to check-immediately following the weekend after the settlement meeting, when I was talking to my colleagues. I think I explained to the Treasury Committee that there were actually two mistakes. There was one financial error, and there was a mistake made in not reporting the case instantly to our high-risk corporate programme board.

Q242 Mr Bacon: When you say not reporting the case, do you mean not reporting the settlement?

Dave Hartnett: Not reporting the settlement, yes.

Q243 Mr Bacon: So you cut a deal but you did not tell anybody? Is that what you are saying?

Dave Hartnett: No, no. I am not saying that at all.

Q244 Mr Bacon: I am not trying to be glib-I know it sounded glib; I am often accused of that, I’m afraid-but that sounded to me like what you just said. A deal was struck but it was not reported immediately. What does that mean, if not that you cut a deal but you didn’t tell people?

Dave Hartnett: A settlement was reached, which all the HMRC people believed to be within the authority of the sector lead and the case relationship manager to reach. Over the weekend following the settlement, the sector lead began to realise that it was not within his competence to do that. That was the other mistake, and it was referred late to the high-risk corporate programme.

Q245 Mr Bacon: Right. How late? How much later was it referred?

Dave Hartnett: Days.

Q246 Mr Bacon: And why wasn’t it within his competence?

Dave Hartnett: Because he had not fully consulted the other areas of our business interested in the issues.

Q247 Mr Bacon: That sounds familiar, actually-not consulting people who are familiar with the issues. So that was the second mistake: the deal was not reported immediately. The first mistake was what you described as a financial error.

Dave Hartnett: The NAO described it as such, but I agree with the description.

Q248 Mr Bacon: That was also the legal error in thinking that you could not charge interest, was it?

Dave Hartnett: Yes.

Q249 Mr Bacon: Whereas you actually could. In coming to the conclusion that you could not charge interest, did you consult lawyers about that?

Dave Hartnett: No.

Mr Bacon: You didn’t?

Dave Hartnett: No.

Q250 Mr Bacon: Since it is very common to charge interest-it is standard practice to charge interest-wouldn’t consulting lawyers on that question have been the obvious thing to do?

Dave Hartnett: There are two things to say with that, Mr Bacon, one of which I touched on last week. I had overseen the 2005 settlement of the issues, and I knew of the legal impediment to collecting interest. My colleagues were in a similar position of knowing, and so was someone from the taxpayer. We were all confident that we knew that. The dispute that was in the hands of the lawyers was over the principal sum; it was not a dispute about interest.

Q251 Mr Bacon: Hang on a minute. The legal impediment that you thought you were confident about was a supposed impediment to the charging of interest.

Dave Hartnett: Yes.

Q252 Mr Bacon: But the 2005 settlement that you were talking about was several years previously.

Dave Hartnett: Yes.

Q253 Mr Bacon: Would it not have been standard practice to revisit why it was that you were so confident about the nature of that impediment?

Dave Hartnett: Well, because I had not been running the case, but those with me had, and between us-and I am very sorry about this-we all believed the impediment still to be there.

Q254 Mr Bacon: But you actually wrote to Goldman Sachs-

Dave Hartnett: I don’t think-

Q255 Mr Bacon: Sorry, not you. HMRC wrote to Goldman Sachs and told it that the interest would run-that the interest would continue to roll up and be due if it did not settle with the other 21 companies.

Dave Hartnett: I’m sorry, Mr Bacon, I was unaware of that letter until you mentioned it to me.

Q256 Mr Bacon: I bet you found out afterwards.

Dave Hartnett: I might have.

Q257 Stephen Barclay: At our previous hearing you said that you could not recollect whether a detailed note was taken at the meeting. Have you now had an opportunity to check?

Dave Hartnett: I have not yet had a chance to check, Mr Barclay.

Q258 Stephen Barclay: You have not had a chance to check?

Dave Hartnett: No, not yet, but I will do so.

Q259 Stephen Barclay: When?

Dave Hartnett: Tomorrow, if I can.

Q260 Stephen Barclay: In your evidence to us last week, you said, "We keep very detailed records, indeed." It’s surprising that you felt it necessary to phone the global head of tax at Goldman Sachs to summon him from New York and a meeting record was not kept.

Dave Hartnett: Mr Barclay, I know there is a meeting record, which HMRC has, that was prepared by the taxpayer. I do not know whether there is an HMRC record as well.

Q261 Stephen Barclay: But these were senior people at the meeting. When the mistake was identified on the Monday, was a note circulated to lawyers at that stage?

Dave Hartnett: I spoke to Mr Inglese to seek his advice. I do not know whether he circulated a note; I certainly did not.

Q262 Mr Bacon: When did you speak to Mr Inglese?

Dave Hartnett: I think on the Monday, but I need to check, Mr Bacon. I’m sorry.

Q263 Mr Bacon: You think it was the 22nd, after the Friday.

Dave Hartnett: Very soon after, yes.

Q264 Mr Bacon: On the interest and the nature of the mistake, you prefaced your remarks about the interest by saying that you were involved in the 2005 deal. But it was in October 2005 that HMRC had warned that the interest would continue to accrue. It defeats me how, in apparently being so confident that you could not charge interest and therefore it was not something that you seriously thought about doing, the record showed the exact opposite. You had warned them about the fact that you would continue to charge interest, and yet you, or the people you were with, did not revisit the question. If you had, you would have discovered that there was not an impediment. That is the whole point.

Dave Hartnett: I think, Mr Bacon, that if the NAO is going to look at the matter, it should all come out then, because there is much more detail to come out.

Q265 Stephen Barclay: You also said, in answer to one of the questions last week, "normally we are able to intervene to correct them", as in mistakes, "because they are spotted by the process. I spotted the mistake here." If the mistake was spotted as quickly as the Monday following the Friday meeting, why was it so binding on HMRC that you were not able to rectify it as would be normal practice?

Dave Hartnett: I am not going to be able to answer that in public, Mr Barclay, but we will be able to deal with it through what Mr Morse is going to do. But let me just say this: there were discussions with the taxpayer shortly after we discovered the interest and the governance error.

Q266 Stephen Barclay: But you accept that legally you were not bound by that decision. Legally you could have rectified it.

Dave Hartnett: I received advice that we could regard ourselves as bound by it, or not bound by it.

Q267 Stephen Barclay: In other words, you were not bound by it.

Q268 Mr Bacon: Mr Barclay has put it as perfectly as one could. You have the option of either going back and revisiting it and saying, "Yes, we will charge the interest" or accepting that it was a done deal, and you went with the "It’s a done deal" option, although you did not have to. That is correct, isn’t it?

Dave Hartnett: Mr Bacon, I have just been passed a note by my lawyers to tell me that this is taxpayer confidential and will need to be dealt with in the review of the case rather than now.

Q269 Mr Bacon: I think we probably will not get much further today, and I look forward to the CAG’s further report. It does seem very odd that not only did you choose the second option of accepting it as a done deal when you could have gone back and revisited it and fixed it, but Mr Inglese signed it off in that way, and so did two tax commissioners. That must have resulted in a negative impact on the revenue. The legal advice was that it need not have been that way. That seems very odd.

Dave Hartnett: If I may, one final comment from me: "must have" may not be right. Mr Morse explained other possibilities last time, and there is more to this that will come out.

Q270 Chair: Okay, we are going to come back to it after both the NAO work and our discussions with the Cabinet Secretary. We will probably expect him to give evidence to us anyway, although he was clearly unable to do so this afternoon.

Right, PAYE. I do not know whether you are dealing with this, Mr Hartnett. One issue slightly concerns us. You have stepped up to be acting permanent secretary in the circumstances, which we completely appreciate. Are you able to do all this work of sorting out all the deals that you get involved in, and making sure that the rest of the Department is functioning well?

Dave Hartnett: I am doing my very best to do that, Chair. It is not the first time that I have stepped up. I am Lesley Strathie’s second permanent secretary-I led the Department for a year once before while trying to do other things. I am doing my very best.

Q271 Chair: We might come back to that, because there are some concerns about your cost cutting.

The PAYE story is a better one than when we looked at it six months ago and we welcome that-it is a better place to be in. Looking at it in that context of a welcome improvement, let me briefly deal with the issues that remain. I am concerned that on page 9, in paragraph 26, we are told by the NAO that you have nearly 7 million individual records that are awaiting reconciliation, which is still 15%, which is one heck of a lot. When will you have done that reconciliation, and what is the impact of those figures on the financial loss to the taxpayer?

Dave Hartnett: I’ll ask Stephen to pick that up, if I may.

Stephen Banyard: Under the old pay-as-you-earn system we had to manually reconcile 16 million or 17 million records a year; under the new one, we expect to have to reconcile 3 million to 4 million a year, so there is a huge reduction in the amount of manual work we have to do. The 6.7 million is the number of cases that needed to be looked at by our staff this year. We are 39% into that and we expect to have completed it this year-on schedule.

Q272 Chair: In this financial year?

Stephen Banyard: Yes.

Q273 Chair: Can you update the Committee on the total income tax forgone because of the delays and difficulties in introducing NPS? Have you got that figure or can you put it together for us?

Dave Hartnett: I have got some numbers for you. For the years 2004-05 and 2005-6, we have only a notional figure. The Committee may remember that when Lesley Strathie was here she explained that notional meant the largest amount, but that we would not have necessarily collected it all. It is important to realise that. That was a figure was £150 million. In 2006-07, the amount forgone was of the order of £100 million.

Q274 Chair: We’ve got £500 million.

Dave Hartnett: Again, I think that is the notional figure. Lesley was at pains to explain that the actual amount that we might have collected in the summer of 2010 stood at £100 million, and by November 2010 it was down to £25 million.

Q275 Chair: What do you in the NAO feel more comfortable with looking at? There is a heck of a difference between £500 million and £100 million.

John Thorpe: Both the figures are included in paragraph 3.11 of the report. The £500 million was the notional amount at the time, in 2006-7; the £100 million was based on the Department’s assessment.

Q276 Chair: Whose notional figure at the time? HMRC’s or yours?

John Thorpe: It is from HMRC records-from the database of open cases. The £100 million is the estimate of what could have been collected at the start of the recovery exercise in June 2010.

Q277 Chair: So what could have been collected is different from what was owed.

John Thorpe: Yes, because had HMRC acted at the time, there was a potential to collect £500 million.

Q278 Chair: Okay, so let’s take £500 million. So far, we are at £650 million-go on.

Dave Hartnett: There is a real problem with £500 million, because that is not what could have been collected; the £100 million is what could have been collected.

Chair: No-yes. It is a question of what could and what should-what was owing and what could have been collected-there is a difference.

Q279 Ian Swales: For example, that includes your changing the de minimis rules, I believe.

Dave Hartnett: I am coming to that next, Mr Swales.

The £500 million includes every penny-

Q280 Chair: That should have been paid.

Dave Hartnett: That should have been paid, but not every penny could have been collected. It would have been cost-inefficient to do so.

Q281 Chair: I accept that. I understand that. Of course, in all these figures we do not have the-

Dave Hartnett: The £300 threshold we introduced for 2007-08, 2008-09 and 2009-10, involved the sum of £266 million.

Q282 Chair: Right. I have another figure of £136 million of underpayments on the 2007-08 cases, not worked or collected.

Stephen Banyard: That is down to £126 million now.

Q283 Chair: That is down to £126 million from £136 million. Then I have another figure of £41 million for recovery of tax forgone following successful taxpayer claims of remission under extra-statutory concessions.

Dave Hartnett: That is correct.

Q284 Chair: Then we have an unquantified amount, which you might like to quantify, of tax forgone from some 250,000 pensioners, where their tax codes failed to reflect their state pension between 2008 and 2010.

Dave Hartnett: We do not have a figure for that yet. I think that is correct, Stephen?

Q285 Chair: Do you have a ballpark figure?

Stephen Banyard: All of these estimates are to an extent unreliable, but the figure for the pensioners’ tax is so unreliable that neither we nor the NAO could put it into the report.

Q286 Chair: Presumably, there will be another figure for these 6.7 million accounts that you have yet to reconcile, of which you have done 30-something per cent.

Stephen Banyard: No, we expect to put all of those into charge, because they are in time. They are for 2008-09 and 2009-10, and we are working them so that they will be in time.

Q287 Chair: So, we are talking about, in total, £1.2 billion or thereabouts.

Dave Hartnett: If the notional figure for 2006-07 is taken, it comes to that order.

Chair: That is a shocking loss to the taxpayer.

Q288 James Wharton: Do you think that, given where we are now and what we now know of how things have progressed, it was the right decision to increase the de minimis limit in order to streamline administration, but obviously taking on the cost of that?

Dave Hartnett: I will bring in Stephen in a moment. We would have had real difficulty if we had not done that. It was a case of managing the work to make sure that we got the work done before the time limits ran out and that we caused less distress to taxpayers than we might otherwise have done.

Stephen Banyard: What the Department faced, as it brought in the NPS system late, was a concertina of work. You cannot carry that bow wave forward for ever; you have to unwind it and start to operate the tax system in real time. By introducing the £300 tolerance, the Department cut out 40% of the adjustments, but only 8% of the yield. It was on that judgment that we took that decision.

Q289 James Wharton: One of my concerns is related to discussions that I have had with HMRC employees who are my constituents about how they work. One of the issues that has repeatedly been raised with me is the lack of flexibility when data are input into systems. For example, when they enter things from people’s tax returns, even if they know a figure is wrong, in many cases they have to enter that figure as it appears on the tax return in order that it can be picked up and corrected later.

My concerns are twofold. First, can you explain-if possible, I would be grateful if somebody could send me the guidance that HMRC employees use-whether, if they have a figure that is wrong, they know within what bounds they are allowed to change it or to raise such a matter at the point of input? If there is some guidance on that, I would like to see it.

My other concern is that a judgment was made that, because there was a backlog of errors, the value of which added up disproportionately to the work that it would take to chase them up, the Revenue decided not to chase people beyond that de minimis limit. If what these constituents are telling me is true-that errors are still being input as data go forward, with the idea being that they are picked up later-is there a danger that this will happen again, perhaps on a smaller scale, because where the systems are not in place early enough in the data input process, we create errors that could have been resolved? Therefore, we may find ourselves making such a judgment again, saying, "Well, we have all of these errors backed up, so we will increase the limit. It doesn’t cost that much". However, we would be building in a system in which errors are commonplace-I am sorry that that is a long question.

Stephen Banyard: The general answer is that I am very mindful of the fact that data quality is extremely important, and therefore, with pay-as-you-earn and the RTI programme, we focused on getting data quality built in from the start. In relation to your point-I know our staff are concerned about this matter-self-assessment is a "process now, check later" regime. Just because the figures have been put in does not mean to say that we have shut our eyes to them. The people who do the checking then receive a list of exception cases, or cases where there is reason for concern, and we can then take a risk choice on which ones we decide to pursue. We have resource to pursue compliance, and we have to use it to best effect.

Q290 James Wharton: The way that the process is working is at the heart of my concern. I am being told by people-and you have acknowledged that staff are raising this with you internally-that they are putting in data that they know are wrong, and people feel that they do not have sufficiently clear guidance or flexibility to correct that at the input stage. I understand what you are saying, which is that you put the data in and then check them. You have just said now, however, that those that stand out get checked, but how many are being missed? And, when you take the judgment on which are worth pursuing, with how many are you saying, "We won’t pursue that", but had the data been put in right in the first place, it would not even have been an issue? I am concerned that-although on a much smaller scale-on a regular basis, the problem that we have seen with this process is actually being built into the system.

Dave Hartnett: Mr Wharton, may I help? We recognise this issue. Our staff have raised it with us, certainly in the past. The extent of electronic filing now, particularly for self-assessment and by employers, and the electronic checks that happen there, throw things out much faster when they are evidently wrong. Those are put on to work lists and we address them. With nearly 80% of individual tax returns and a huge proportion of employer returns now filed electronically, we are detecting these things automatically. It does not excuse the past, however.

Q291 James Wharton: Do you believe that your system, as it is at the moment-accepting reasonable errors and problems that will be thrown up-is working sufficiently well, and therefore, you will not be revisiting it, saying, "We are either going to increase the de minimis limit", or "We are not going to chase those and actually try to get that money in, because there are so many of them, or because there is such a backlog"? Do you believe it works sufficiently well to avoid being in this position again via a different means?

Dave Hartnett: We think there has been a sufficiently significant improvement for us to be able to reduce the de minimis back from £300 to £50. Certainly, on current performance levels, we do not expect to have that problem, but we are watching it very carefully.

Q292 James Wharton: That was my last question, but for clarification, if you could send us a copy of the guidance that staff are given on when they can amend at the data input step, that would be much appreciated.

Dave Hartnett: Of course.

Q293 Chair: I have two other areas that I want to pursue-I am looking around to see if anyone else wants to. One is that on figure 9 on page 36, you have now calculated that you have £1.9 billion in overpayments by individuals, and £1.1 billion in underpayments by individuals. Am I reading that right?

Dave Hartnett: That’s £1.1 billion-I can’t find it.

Chair: Is that £1.1 billion by individuals?

Dave Hartnett: By individuals.

Q294 Chair: How confident are you that you will get that money in? What proportion of it?

Stephen Banyard: We are fairly confident.

Q295 Chair: What-100%?

Stephen Banyard: Yes. Within the normal collection parameters, of some people going missing, some dying, and some becoming insolvent-within those sorts of limits. These are within time limits, and we would expect to collect them. The way that the pay-as-you-earn system works is that at the end of the year, most people-85%-will have paid the right amount of tax. Some will not, for very good reasons. Usually, it is because the taxable benefit that they get in the year has changed. We make a correction at the end of the year and we collect the money normally in the following year.

Q296 Chair: You do not normally have this size of correction, do you? Over £1 billion a year correction.

Stephen Banyard: It is not £1.9 billion-

Chair: Overpayments-which you have to pay back.

Stephen Banyard: That is cash.

Chair: That is cash, but £1.1 million underpayments-is that the sort of size we get every year?

Stephen Banyard: That is cash too. £1.1 million underpayments, can I-

John Thorpe: These issues cover two years, of course, because the Department had to reconcile two years together-

Chair: And they have not completed the reconciliation-there is another 15% to come.

John Thorpe: Yes, so it captures two years.

Q297 Chair: But so far? What is the size usually of the underpayments?

John Thorpe: I do not have a number for that.

Q298 Chair: Do you?

Stephen Banyard: For?

Chair: Underpayments.

Stephen Banyard: Yes. For this year or last year?

Chair: Usually.

Stephen Banyard: Usually. We are operating a new system but last year the overpayments were 5.6 million (customers). We have completed the overpayments-

Q299 Chair: This is 2011-12? Sorry, 2010-11.

Stephen Banyard: For 2010-11 the number of overpayments is half what it was last year. It was 2.3 million (customers with) overpayments.

As for the underpayments, we are about to start issuing. We estimate-I emphasise "estimate" because, until we have worked through all the records, we will not know the position-that on a like-for-like basis, compared with last year and with the £300 limit, there will be half as many. But, because we have lowered the limit from £300 to £50, there will be the same number. So there will be about 1.2 million-

Q300 Chair: Billion-we are talking billions here.

Stephen Banyard: 1.2 million adjustments, and the estimated size of the adjustment is about a half of what it was last year.

Chair: Oh yes, I see what you mean. Anne.

Q301 Mrs McGuire: May I move us on to some of the work that you are currently preparing for, which is the linkages between the child benefit and the new systems, and indeed the universal credit? I think I saw a headline recently-but I never believe everything I read in the newspapers-that indicated that the Treasury effectively thought that the introduction of the universal credit was just not achievable, certainly within the time scale and the financial limits that had already been set. Would anyone like to proffer an opinion on that particular comment? Was it just a bit of fluff in the newspapers? Was there a genuine dispute between DWP, the Treasury and indeed HMRC, or were you just piggy in the middle between two major Departments?

Dave Hartnett: Stephen is closer, as the SRO for real time information, but both Stephen and I sit on a ministerial committee dealing with this, and I am not sure I would call it a bit of fluff in the media but, certainly, we did not recognise from the discussions that we had been involved in the descriptions that appeared.

Stephen Banyard: We cannot comment on the universal credit programme because that is for DWP to comment on. What I can say is that I-

Mrs McGuire: But you are managing it.

Stephen Banyard: I am responsible; I am accountable for the real time information programme, which provides the-

Q302 Mrs McGuire: So you are crucial to the delivery of the universal credit.

Stephen Banyard: Absolutely.

Q303 Mrs McGuire: Because if your information is wrong, then lots of people out there are going to find themselves in dire financial difficulties.

Stephen Banyard: And we are very mindful of that.

Q304 Mrs McGuire: I do not know whether or not the child tax credit legacy still looms hard and you have a big sign up saying, "Please do not repeat what happened with the child tax credit situation."

Stephen Banyard: We will put every emphasis that we can, which we are doing, on collecting and transmitting accurate data-

Q305 Chair: You’re not doing anything on real-time information at the moment. That is one of the things that the Report says. The Report says that, because you have been trying to sort out both the PAYE numbers that we all get-whatever you call it-and the returns, you have done nothing on real-time information.

Stephen Banyard: No, that is incorrect, Chair. We have a major programme running. I do not know where that comes from-

Chair: My reading of the Report.

Stephen Banyard: But we have a substantial programme running on RTI, and I can reassure you-

Chair: You might have a policy, but your guys are busy pursuing 2008-09 and 2009-10.

Q306 Mrs McGuire: Can I ask the NAO why? Are we misinterpreting what is in the Report?

Stephen Banyard: Yes.

Q307 Matthew Hancock: Which bit of the Report?

John Thorpe: The commentary on RTI is in paragraphs 3.38 to 3.41. We describe the programme and the time line in terms of how the Department is implementing RTI, so we just comment on the principal milestones. Our commentary is principally around the challenges around data quality and how those are being addressed. My understanding is that it is a separate programme from the work that is going on around the NPS stabilisation.

Q308 Ian Swales: It says in 3.41 that the pilot phase runs from "July to October"-so it should be completed in two weeks’ time-"prior to full implementation from November 2011." So how are we doing?

Stephen Banyard: Can I help? The first thing to say is that we have done a lot on the new pay-as-you-earn system to improve data quality, and you will have read in the Report that the Department has repaired 11 million records. They were not wrong, but we needed to check to make sure. As a result of that, the tax codes that we issued this year-the Committee will be concerned with this-were very accurate compared with the inaccurate codes from the year before. We achieved an accuracy of over 98%.

Another key indicator, and one that DWP are interested in, is how well the data coming in from employers matches into our systems, and whether it will drop into the right customer account. We have been working on that over the past two or three years, and we have increased its accuracy, in successive years, from 97.3% to 97.8% to 98.3%. The acceptance criterion for universal credit is 98%, so we have already achieved what universal credit needs from us. We are not content with that, and we have set up a separate data quality programme within the RTI programme.

We are particularly looking to find out where the problems are arising, and we have developed a tool that enables us to look by employer to see where the problems are. Over the past three months, we have piloted-this is what you are reading in the Report-the introduction of customer relationship managers, who will go out and work with employers and talk to them about data quality and its importance.

Most employers need good quality data, just as we do, and they very much understand our need for it, but different organisations have different data needs. For example, you may work for the House of Commons and they have you down as Mrs A McGuire, but on our records you would be Mrs Anne McGuire, and something as simple as that can cause problems. We are working through with employers. We have a list of the employers who are likely to cause us most problems and we are working with them in a collaborative way to improve our and their data.

Q309 Mrs McGuire: The thing that identifies me within the system is my national insurance number, so why do I even need to have my name there at all? I do not want to put it into the public domain at the moment, but I can give you my national insurance number off the top of my head-I have had it for a long time.

Stephen Banyard: And you are wonderful, but not everybody can do that.

Q310 Mrs McGuire: All right. I am being a bit facetious here, but people do carry their national insurance numbers, and, frankly, I never cease to be amazed-I do not know about my colleagues-that when you ask a constituent for their national insurance number, they can just run it off the top of their head. Many people do know it. Is that not the identifier, and not whether I am Mrs Anne or Mrs A-or, indeed, Mrs A McQuire, which I sometimes get and which is even more confusing?

Dave Hartnett: It is the primary indicator, and many people do exactly what you do, Mrs McGuire. They know their national insurance number off by heart. Others no longer know their national insurance number. Others still have been given temporary national insurance numbers at some stage. You begin to see how the confusion can arise. Our challenge is to cleanse all that.

Q311 Ian Swales: Just on this very point, in data quality terms, you talked about an acceptance criteria of 98%, so that is 2% error. How many errors would that be if you hit your acceptance criteria? A million?

Stephen Banyard: The acceptance criteria are around data hitting the account first time. We then try to match it using our staff. The matching criteria that we use are normally name, NI number, address, date of birth and gender. It is important that we can match accurately. While we are not content with 98%, it is a good starting point.

Q312 Mrs McGuire: Can I come back on the second part of my question, which was about the child benefit changes? What do you perceive to be the major challenges in implementing a real-time database to assess whether an individual or family are entitled to child benefit, under the new system that is coming in?

Dave Hartnett: I think, Mrs McGuire, that the starting point is this: we have got to make system changes to deal with that, and we have until February or March next year to understand what those changes are-once the policy design is completed, which is in the hands of our colleagues in the Treasury. It is hard to answer your question today. We do not envisage any special difficulties with real-time information and child benefit.

Q313 Mrs McGuire: That does not strike me as a long read-in time to introduce something that will make a significant difference.

Dave Hartnett: No, we have been thinking about it.

Q314 Mrs McGuire: You’ve been thinking about it.

Dave Hartnett: We have been thinking about it for a long time. I hope we understand what the options might be, but there are still decisions to be made.

Q315 Chair: But, to be clear, you have done no in-year changes. Para 23, page 8 of the NAO Report says you have put the resources into the recovery work. It states: "This recovery work was extensive, covering over 11 million records…A consequence of diverting operational resources to recovery is that some in-year changes to individuals’ records have not been processed".

Dave Hartnett: But they are in-year changes to individuals’ pay-as-you-earn records, which, currently, will have nothing to do with child benefit.

Q316 Chair: But they could have something to do with their tax. There may be overpayments or underpayments.

Dave Hartnett: Indeed.

Q317 Chair: So it will add to your overpayments-underpayments problem.

Stephen Banyard: Nobody will pay the wrong tax at the end of the year.

Q318 Chair: Why? Because you will do in-year changes? This says that you are not.

Stephen Banyard: We want to do in-year changes.

Q319 Chair: But you are not.

Stephen Banyard: We are in some cases. May I give some context? The position we have faced is this concertina of work at the start. In-year information comes to us from three sources: people write to us-some 9.5 million letters a year-around 16 million people telephone us, and the computer throws up items for us to look at. Perhaps I should say this: 12 months ago we had a backlog of post, which was noted by the Committee. We are fully up to date with our post here. We have been working on it over the summer. Our postal position is up to date.

Q320 Chair: What is your definition of "up to date", out of interest?

Stephen Banyard: We have only five days’ post on hand.

Q321 Chair: So do you reply within five days or do you open it after five days?

Stephen Banyard: No. People will start replying to a letter, if it is urgent, at the five-day point. It will go out then.

Q322 Chair: When will someone get a letter? That is what I am really after. That is what we, and the punters, care about.

Stephen Banyard: You ought to get a letter well within the service standard period.

Q323 Chair: Which is what?

Stephen Banyard: Three weeks.

Q324 Chair: Right. Do you think that you are there?

Stephen Banyard: We want to do better than that, but we have got to that position. People will begin to see that in the autumn, because we have only just got there. We have also improved the telephone service. What we are left with are the work-management items, which are quite a mixed bag. Some of them are about data quality, which we have been talking about.

For example, we may have a record that matches on four of the five items, but perhaps a character in the name is wrong. We will match the record and throw up a work-management item and ask whether it is right or wrong. It may be a compliance check, as we discussed earlier, such as a repayment. The system will ask whether we want to check it. We set a level or criteria for checking.

A lot of those work-management items are about information management. Some of those items relate to taxable benefits, such as company-paid health insurance or company cars, that have changed during the year-we may not have worked through the items that come from the computer this year, so we will pick them up at the end of the year-but that is the final piece of the jigsaw of trying to get pay-as-you-earn back up to date. We are carrying out a major review to address those work-management items and to determine which ones we actually need.

Q325 Mrs McGuire: My final question is about the integration of child benefits into the real-time programme.

Say that I am a school teacher with three children and that I get promoted in August, which takes me over the threshold. When will you catch up with me-if I may put it in those terms-and withdraw my child benefit?

I bear the scars of the child tax credit, which left people with horrendous bills. I remember someone, although they were perhaps not in your position, Mr Hartnett, explaining that there is a difficulty with the culture of the HMRC because it is only used to collecting money, rather than giving money out. What is in front of those constituents who, during the tax year, flip over the point at which they are no longer entitled to child benefit? Potentially, significant sums of money are going to be withdrawn.

Dave Hartnett: I agree. If we know that the child benefit needs to be withdrawn, there will be an in-year change; if we do not know, it will be dealt with in the following year.

Q326 Chair: You won’t force people to pay it back? Who loses out in Anne’s example of a woman being promoted in September without the local authority telling you until the following April?

Dave Hartnett: If the woman does not contact us to tell us of the change, there will be a tax charge for the year of promotion, but, where it is appropriate to do so, we will encourage people to take themselves out of child benefit.

Chair: That is very unclear. Anne’s example is quite good.

Q327 Stella Creasy: Flipping it round the other way: what about the person who breaks up with their partner and becomes entitled to child benefit? How will you deal with that?

Dave Hartnett: If we know that has happened, and if the person has come out of child benefit, we will put them back into it.

Q328 Stella Creasy: How quickly can you operate that system? It sounds as if you will not make an adjustment until the following tax year for people who are no longer entitled to child benefit. How quickly will you be able to assess whether someone is entitled to child benefit?

Dave Hartnett: If someone contacts us, we will be able to make such an assessment in a very short time.

Q329 Stella Creasy: What’s a short time?

Dave Hartnett: Weeks at the maximum; perhaps even immediately.

Q330 Stella Creasy: So how can you do it if they are entitled, but it will take longer if they are not entitled, therefore?

Dave Hartnett: Because we’re more likely to find out if someone is entitled than if they are not entitled. I think that’s how it will happen.

Q331 Matthew Hancock: May I push you a bit more on real-time and the change programme around that? You have described how you are catching up with the backlog within the existing system and how you are improving the data, to go up from 97.3% to 98.3%. Operationally, it is probably better described as a 2.7% error rate going down to 1.7%, because one is much bigger, but it’s the proportion that matters. But there are other elements, particularly the link to the new universal credit, that are part of the change programme. How is that work proceeding?

Stephen Banyard: The RTI programme is on track and we are confident that we can deliver on time. We have hit all of our recent milestones. We are running a 12-month live trial, and we have run workshops to start preparing 320 volunteer employers. We are on track to start the pilot in April, and it will run for 12 months, which will enable us to test the system. If-

Q332 Matthew Hancock: Excuse me. Are the 320 people all in work?

Stephen Banyard: I said 320 employers, representing more than 1 million taxpayers.

Q333 Matthew Hancock: Do you also include unemployed people in that trial?

Stephen Banyard: This is a trial of the operation of pay-as-you-earn.

Q334 Matthew Hancock: The real-time pay-as-you-earn, rather than the DWP’s end of it?

Stephen Banyard: Yes. HMRC is responsible for delivering real-time pay-as-you-earn information.

Q335 Matthew Hancock: But you are also jointly responsible for delivering a real-time link-up with the universal credit.

Stephen Banyard: Indeed, and we will provide the information to them through that linkage.

Q336 Matthew Hancock: And the milestones on that interface?

Stephen Banyard: Yes, we are building that, and we are on time with all our IT.

Q337 Matthew Hancock: How would you describe the working relationship with the DWP team that is doing the same?

Stephen Banyard: First class.

Matthew Hancock: I thought you might say that.

Stephen Banyard: I would, and it is. Most of HMRC’s programmes require us to work with a lot of stakeholders. The programmes that I have been involved in require me to work with tax agents, software manufacturers, employers and various groups. You don’t deliver programmes unless you work well-or try to work well-with those groups.

You need to do three things: first, share goals-you need to know what you are trying to do and agree it. Secondly, you need good working mechanisms, such as steering groups. We have cross-representation, we have seconded people across each way and we meet regularly. Thirdly, you need good personal relationships and you need to trust each other. We have, from the outset, set out to build good personal relationships with the DWP team. Relationships are good.

Q338 Ian Swales: Without wanting to go back over the ground that Mr Wharton covered earlier, we start from a position where anecdotally, your staff are not even allowed to, when they are inputting data, take information that they can see in front of them and use it to get the data right. If we are expecting to move to a situation where all these benefits and credits come together-my question is about the management and governance of it, not just the systems.

Taking the points that Mrs McGuire and Ms Creasy were making, it ought to be automatic in the future that you know when someone has reached a certain point in the tax system and then benefits change, and vice versa. If it isn’t automatic, this whole new system is not going to work. My question is not just about whether the systems will do it, but what you are going to do about management, governance and control. Who is going to make the decisions and communicate these changes?

Stephen Banyard: Can I take the child benefit part of the question first? Real-time information will help us on that, because even if you don’t phone us up and tell us, we will spot that your income on a monthly rate has gone up and may have tripped the higher rate boundary. We would have governance around those sorts of things, perhaps to go in and ask the employer or you, "Have you crossed the boundary?"

Q339 Ian Swales: But that’s exactly the point. It all sounds extremely woolly and gentlemanly. The point is that something has changed and you know it has changed. Aren’t you going to do something to avoid overpayments and underpayments? "We might perhaps let you know or talk to your employer" doesn’t give me the warm feeling that the new, dynamic real-time system is going to have the right management structure around it.

Stephen Banyard: I assure you that it will. We haven’t operated a real-time system. We have to get used to it.

Mrs McGuire: That’s what worrying and concerning us.

Q340 Stella Creasy: It’s the postbag we’re going to get from the family that breaks up, but which may still be under the same roof for a period or can’t prove to you that they have broken up; or perhaps from a person who gets promoted but doesn’t get their salary straight away. All those people will come to us in our surgeries and will be looking to us to help them sort out what’s going on with you guys.

Stephen Banyard: I understand that, but those are not ‘hard’ facts. The real-time programme, if you bring that into it, gives us a facility to be able to spot those cases and intervene.

Q341 Ian Swales: I don’t necessarily expect you to read people’s minds or understand everyone’s social situation, but you have to be able to see data moving on tax and credit. Otherwise, the universal credit system isn’t going to work. We need to have a system that works. Are we going to trip up over the fact that more than one Ministry is involved, for example, or is that all going to get sorted?

Stephen Banyard: I hope we do not trip up over two Ministries being involved. We have put very good governance, which has been externally examined, around the programmes. As we move into implementation, that governance will have to adapt as we face new situations.

Q342 Ian Swales: It is the operational governance I am talking about. I am not minimising the issue of getting the systems in place-we all know that is going to be massive. But assuming success, are we confident that we will have operational governance that is fast and avoids the situations we have been talking about?

Stephen Banyard: There will be and I accept the point. In a real-time world we will need to respond in real time to those sorts of situations.

Q343 Stephen Barclay: This really follows on from Mr Swales’s questions. Of the 54 projects in the change programme, how many have started?

Dave Hartnett: Twenty-four have got-

Q344 Stephen Barclay: You are working on the business plans for 24, so they have not really started, have they? It is just the initial work.

Dave Hartnett: No, they have started, Mr Barclay. John may be able to help.

Q345 Stephen Barclay: It says, "preparing full business cases for the 24 projects to deliver the £964 million." It is a really simple question. We have 54 projects in the change programme.

John Keelty: I cannot give you the number immediately, although it may come. For example, on some of the reinvestment of the savings, we have already started recruiting people into various areas so that we can start those compliance projects going. We have already started to bring in more yield as a result of the reinvestment of the savings we are making this year. There are a number of projects that we have started this year. It is not the case that we are just planning. Yes, we are planning for some that will materialise in 2013-14 and beyond, but for those that we need to get going now, we are well on the way to doing that.

Q346 Stephen Barclay: Sure, but that was not really my question. I appreciate that for 24 you are working out a business case, but what I am saying is that by 2015 you have to deliver 54 change programmes. That seems quite a lot for a department that has business as usual and has other issues to address. Perhaps you could send us a note about how many of those have started. How many of those that have started are off-track?

John Keelty: Again, I am not aware of any but I do not have the precise details. We monitor these reports on a monthly basis, so what we get each month from them is to see their current status. As we work through them, there will be some that will be rated "green"-absolutely perfect, no problems-and there will be some that will be rated "amber", where we need to do more work. I have run programmes before and that is what happens.

Q347 Stephen Barclay: I fully appreciate that you have ratings; hence my question. Surely senior management has a sense, if you monitor it monthly, of how many of these 54 change programmes have started. We have less than four years to deliver them-it seems a lot of work-and it seems odd if senior management does not even know how many have started and how many are already behind schedule.

John Keelty: We have just undertaken a mid-year review of the change programme and the conclusion we got from that is that we do not need any radical redesign or radical changes. Everything is moving in the way that we would expect it to move at this point in time. That does not answer your numbers question, but senior management at this time have seen that mid-year review and are comfortable that it is progressing in the way that we would expect it to progress.

Q348 Stephen Barclay: Sure, but a 25% cost saving and those 24 projects are just under £1 billion, which suggests more than a bit of tinkering and modification. Perhaps we could look at external spend. Fifty-four change programmes suggests that you may need some external resource. I am very conscious that the Cabinet Office has strict controls around external spend. How much are you budgeting this year to spend on external resource for these change programmes?

John Keelty: For consultancy we are budgeting about £5 million.

Q349 Stephen Barclay: So £5 million on consultancy. What about other professional support?

John Keelty: We have, of course, outsourced our IT development. So that is also being spent outside. We have outsourced our accommodation so we have quite large amounts that go out through that.

Q350 Stephen Barclay: Okay. So legal spend, that sort of thing, do you have a total budget in mind to deliver? I am asking specifically about the change programmes because again, in previous hearings Mrs McGuire and Mr Swales talked about staff morale in HMRC. There is always the question around how much you spend on staff internally for the tax that you recover. There is a trade-off in terms of how much spend and what that brings in. As part of this change programme, which is to save money, what is the total quantum in terms of professional external support?

John Keelty: I’m sorry; I don’t have that figure.

Q351 Chair: I think you’ve got to provide us with that figure. It’s really important and it is an issue that Mr Barclay has raised, so you might have thought about it before you came, because we have raised it every time.

John Keelty: Of course.

Q352 Stephen Barclay: Again, this is not a new issue to the PAC, but another area is about staff moving mid-project and whether that links into deliverables. Can you assure us that for the 54 projects, all staff will have interim milestones and not be able to move prior to their milestones? I am not suggesting that staff should be in place until 2015. That is clearly not a sustainable position, because you need to keep people fresh and motivated. Can you at least assure the Committee that no staff moves for those running the projects will happen that are not linked to a deliverable milestone?

Dave Hartnett: Let me pick that up, Mr Barclay. We don’t aim for our programme and project managers to move, but some will. It is in the nature of a huge organisation such as ours that some of our very best people are managing those programmes and projects, and part of the task we have given them is to develop those behind them.

Q353 Chair: I think our interest as a Committee is, do you aim to keep them there? The most frustrating thing we see as a Committee is that lack of responsibility and accountability. You come back three years later and somebody else is there. It is not that you aim for them to leave, but do you aim to keep them there?

Dave Hartnett: We certainly do.

Q354 Chair: Do you aim to keep them there?

Dave Hartnett: We plan to keep them there, but all sorts of things can happen in those four years. It is perhaps worth adding that the major projects authority in the Cabinet Office has recently sponsored a starting gate review of our change programme, which has had many good things to say about the governance and how tightly we are managing it. I wanted to say, Mr Barclay, that you asked us how many were off track. I cannot remember the technical term for them, so can I call them "hit squads"? They go in when things go off track. To my knowledge-others may need to correct me-we have only done that twice and on some projects we are ahead of plan at the moment.

Q355 Stephen Barclay: I wasn’t suggesting that you could stop people leaving.

Dave Hartnett: I understand that.

Q356 Stephen Barclay: I am saying that for example the FiReControl project had five programme directors over six years, and a number of those moves were within the civil service. Interim milestones should be set, which are linked to people’s appraisals and therefore to people moving. A thing that we repeatedly see on the Committee is that that is not being done. I do not get an assurance from your answer that that has been done in this case, but perhaps that could be taken away. Who is the senior responsible owner for the 54 projects?

Dave Hartnett: Lesley Strathie.

Q357 Stephen Barclay: And she’s not in post at the moment. She’s off, is she not?

Dave Hartnett: She is off at the minute and Mike Clasper has filled that role while she has been away.

Q358 Stephen Barclay: Right, and she will resume that as SRO.

Dave Hartnett: Yes, that’s the plan.

Q359 Stephen Barclay: And how will she, given her other duties, get oversight of 54 projects?

Dave Hartnett: We have a core project team, which is stronger probably than we have ever put into any programme before, and the starting gate review confirmed that. Our CIO is the change director. He and his team are leading the oversight.

Q360 Stephen Barclay: So if she is personally responsible, which an SRO is, how many of the 54 projects would need to go wrong for her to be held seriously responsible from a career perspective?

Dave Hartnett: I don’t think I know the answer to that, Mr Barclay, largely because we do not plan for any of them to go seriously wrong. One of the key messages from the recent review is that the governance is in really good shape and would identify things like that, and we would act quickly.

Q361 Stephen Barclay: Should you not plan? It is the nature of projects that if you are running 54 projects, some will go wrong. We have two PAC hearings a week because so many projects, sadly, go wrong. It seems to be a leap of blind faith to assume that all 54 projects will go well.

Dave Hartnett: Not at all. One thing we have learned time and again from the NAO is that it is all too easy to be over-optimistic and that is exactly what we are trying not to do.

Q362 Chair: Can I come in on that? Following on from that, you have no contingency at all. You were set a financial cuts target. You are working entirely on the assumption that you will achieve them all with absolutely no contingency. Is that wise?

Dave Hartnett: Chair, we have no cash contingency. Given the state of public funding-

Q363 Chair: You’ve got to achieve your cuts. You have no contingency plans. Of course, you must achieve the financial cuts; that is an imperative from the Treasury. You have no contingency plans for alternative ways in which to achieve those financial cuts other than the 54 projects and the other bit.

Dave Hartnett: No, since the NAO review, and during it, we have developed other plans as to what we might do if we needed a contingency.

Q364 Chair: They are not in the NAO Report, are they?

Dave Hartnett: No, because our thinking on this is very recent. It was prompted by the NAO work.

Q365 Chair: So, you are creating contingency plans?

Dave Hartnett: May I quickly tell you what we are doing? For example, we could speed up the reduction of the number of buildings that we occupy. We can speed up our process change in our tax systems and in our PaceSetter work as well. Those are the sorts of areas in which we would find contingency. It is important to say that we have a proud record of delivering the savings that we set out.

Q366 Chair: No, you don’t. We looked at that. The very first Report that we looked at across Government was the performance of all Departments in relation to the 5% year-on-year cuts in the last spending review. I’m looking for the figure. Perhaps the NAO can help me. On page 16, paragraph 1.15, it says that 12.6% of savings were not evidenced, not new and not reported net of ongoing costs. So, you don’t have a proud record.

John Keelty: That review took place at the halfway point when we had planned only to realise something like 40% of the savings.

Q367 Chair: What that review showed was the savings that you had reported. It was not showing the savings that you hadn’t reported. These were savings you had reported and 12.6% of them were not evidenced, not new and not reported net of ongoing costs.

John Keelty: By the criteria, that is absolutely right, but this was halfway through that SRO7 efficiency programme. We learned from that.

Q368 Chair: Well, I rather resent you saying, "We have a proud record." The last time we looked at this we found that our 12.5% were again not evidenced.

John Keelty: Some 87% of them were sustainable efficiency savings.

Chair: You should not be putting up savings and pretending that there are savings when they are not evidenced. That is not how any public servant in any Department should be operating.

Q369 Mr Bacon: Mr Hartnett, you have said that, prompted by the NAO, you are now doing something about the contingency. Why did it require the prompting of the NAO to do something about the contingency?

Dave Hartnett: We were actually thinking about where we would have back-up for our cost-savings project if any started to go wrong. The engagement with the NAO on the cost-saving Report was significant. It made a number of recommendations.

Q370 Mr Bacon: And one of them was about contingency?

Dave Hartnett: One of them was an observation about contingency.

Q371 Mr Bacon: What interests me is that the failure to have contingency is a hardy perennial. Among the many mistakes that the Rural Payments Agency made was the failure to have a contingency plan. Actually, in that particular case, it had one and then it scrapped it. Such behaviour is commonplace, and we have seen it many times. Folk come to this Committee with projects that have gone wrong and where there was no contingency plan. I would hope that by now tattooed on the eyelids of permanent secretaries are the words, among many others, "Make sure you have a contingency plan." Therefore, why did it require the prompting of the NAO for people to think, "Hey, we need a contingency plan." Why was it not already so deeply embedded in the DNA that the NAO noticed and commented on the fact that there was a robust contingency plan? Why was it not that way round?

Dave Hartnett: Because we wanted to get our change programme going. We knew we had to do this, and the NAO helped us-

Q372 Mr Bacon: You make it sound like because you were in a hurry. The Rural Payments Agency was in a hurry; it was in such a hurry that it chose the shortest of all the available timetables, as well as choosing the most complicated of all the possible payment systems. But the evidence is over many projects that being in a hurry is not enough. In fact, being in a hurry can cause delay.

Dave Hartnett: I agree with that.

Q373 Mr Bacon: But back to my question. Why did it require the NAO to prompt you to have a contingency?

Dave Hartnett: The crucial thing for us, Mr Bacon, was to make sure we understood the risks. That can be done certainly without cash contingency. There was some thought that maybe we should have had some cash contingency, and we could not do it, which is now why we have looked in the other areas that I have mentioned.

Q374 Chair: And 36% of your current proposals’ new cost reductions are in the final year, so if you fail on those, you’ve had it.

Dave Hartnett: If we fail.

Q375 Matthew Hancock: One thing that came out of that is that you mentioned that part of your contingency is to do more on buildings and make greater savings there. I wonder why you are not doing that anyway.

Dave Hartnett: Well, we are.

Q376 Matthew Hancock: Hold on. You cannot be doing the same amount as is in your contingency. You must be doing more.

Dave Hartnett: No. We are doing a lot on buildings. What we can do, Mr Hancock, is to speed it up. Let me tell you the record for last year. I think we achieved 43% of building savings across Government, despite only having 14% of the estate.

Q377 Matthew Hancock: But if you can speed it up, and these are savings that you could deliver, why don’t you deliver them?

Dave Hartnett: Well, we are trying to do that.

Q378 Chair: I have to say that as an Arts Minister I spent for ever trying to get you out of Somerset House. We finally succeeded and put it to better use, but you were very, very, very late in getting out of there.

Mr Bacon: Your fingernail marks are all over the outside of the building.

Dave Hartnett: We are about to mothball it to save our costs.

Q379 Chair: Can I raise two other issues that are of concern to me? One is that you are assuming a cut in sickness absence, from your assumption of 9.9 days in 2010-11 to 7.9 in 2011-12. Did you achieve your 2010-11 target?

Stephen Banyard: We very nearly achieved the sickness absence target last year, or we hit it, and this year we are in front of profile for getting down, so we believe that we will get there.

Q380 Chair: So what was "very nearly"?

Stephen Banyard: I can only speak for my own area. The target there was 10.76 and we achieved 11.01, but we are currently tracking at 8.3.

Q381 Chair: The figure I have got here for sickness absence in 2010-11 is 9.9.

Stephen Banyard: I was speaking for my own area.

Q382 Chair: And by 2011-12 you want to get to 7.9. That is out of the report, isn’t it, Jane?

Jane Wheeler: The target for HMRC was an average of 9 days per person in 2010-11, which the Department did not achieve. I am not sure how far it was below it.

Q383 Chair: And you were at 9.9. So it is not very nearly-10% is not nearly, on these figures.

Stephen Banyard: I gave you the answer for my own area, which is about third of the Department. We made a substantial reduction last year. We did not hit the target, but I am confident that we will beat the target this year.

Q384 Chair: Okay. I understand that you looked at your bit. Mr Keelty, you have some responsibility. You are here answering for this whole change programme, where you do not know where you are on the 54 projects. I picked up that you have assumed a cut in sickness absence. I also picked up that you failed to achieve that in 2010-11. I want to know what you have done in 2011-12, and I want to know how viable your assessment of the cash savings is.

John Keelty: At this point in 2011 for the Department as a whole we are tracking at 8.7 days, and our target is 7.9. As Stephen said, we are on target, we believe, to get the 7.9 by the end of this year.

Q385 Chair: And that will give you your cash savings. So we can hold you, Mr Keelty, to account, if you don’t achieve that. Is that right? Or do we hold Mr Hartnett to account?

Dave Hartnett: I think you hold the accounting officer to account, so that would be Lesley or me.

Q386 Mr Bacon: Mr Keelty, when you say that you’re on target-

Stephen Barclay: You’re not on target.

Mr Bacon: Mr Barclay says you are not on target. Even if Mr Barclay were wrong-and he rarely is-how do you know that you are on target? How do you know who is going to be ill on 12 December or next January? How do you know there is not going to be an outbreak of measles or flu or avian flu?

Stephen Banyard: We don’t. What we do is to look at seasonal patterns. We are trying to reduce sickness absence, so we are looking at where we have now got to. The figures you have been given are the 12 months up to a certain point-today. Those figures are falling fast because the figures in current months are very much lower than they were a year ago.

Q387 Chair: If you know that, why don’t you know what the 2011-12 figure is?

John Keelty: The 2011-12 figure, as of today, is 8.7. That’s the year we are in at the moment.

Q388 Chair: Sorry, why don’t you know the 2010-11 figure?

John Keelty: I believe that is 10.8.

Chair: 2010-11 was 10.8.

Q389 Mrs McGuire: 2011-12 is 8.7. Is that year-on-year or is that from 1 April?

Ian Swales: They are rolling.

John Keelty: These are rolling trends. I have just been passed a note: 2010-11 was 9.65 days, and 2009-10 was 10.49.

Q390 Chair: Do you agree with that?

Jane Wheeler: We have 9.9 days for 2010-11 in the report-paragraph 2.18, final bullet. That is against the average of 7 point-

Q391 Chair: Seven point what? Across Government? Don’t talk about the private sector.

Jane Wheeler: The cross-Government figure was 8.5 for year ending September 2010.

Q392 Stephen Barclay: The reality is that you set a target for 2010-11 and you missed that target. Having set a target, you missed it. You are now about to embark on a major change programme. Do you think a major change programme in isolation will improve or worsen potential sickness rates?

John Keelty: During the beginning part of this year we did a lot around improving the mechanisms and processes for recording sick leave and generally improving it. If we see the tracking for this year, since we introduced that initiative in the Department, the sickness leave has dropped substantially.

Q393 Stephen Barclay: Again, with respect, you are not answering the question I put. My question was not, "Is the trend moving in the right direction?" I would hope, given the focus the Department is placing on sickness, that the trend would be moving. It is pleasing to hear, and the Department deserves praise for achieving that. That is good news.

The point that I think the Chair was driving at is that you set a very aggressive savings target related to sickness, without any contingency, if that is not achieved. Therefore, that saving will have to come from somewhere else if it is not achieved. While past performance is not a guarantor of future performance, given that this target related to a higher target than the eventual one, it should have been easier to achieve than the eventual one.

The point is that, going on your performance to date, you set a target and missed it. Looking ahead, there is going to be a shortfall in the savings. There is no contingency allocated. The question for the Committee must be, "Where else are you going to get that money?"

John Keelty: The failure of the 2010-11 target was not in the spending review period. That is the first thing.

Q394 Stephen Barclay: No, but your starting point is higher.

John Keelty: We are on track this year for the revised target that we set, irrespective of the achievement that we got in 2010-11.

Q395 Stephen Barclay: Sure, but you start from a higher point. So the delivery, the savings, the improvement you need to achieve, are increased, because you are starting from a worse point.

John Keelty: Absolutely, and we are on target to hit those savings.

Q396 Mr Bacon: Are we basically saying that the head count you are planning for is predicated on this particular level of sickness absence, and that is how the saving is achieved? Is that right?

John Keelty: Sorry, I am not sure I understand the question.

Q397 Mr Bacon: If I know that in my organisation I am going to have a 1% sickness absence, I know that I need fewer people to achieve the same amount of work-because only 1% of them will be absent during a calendar year-than would be the case if I knew that 7.9% were going to be absent. Are you basically saying that in doing your sums, you work out how many people you need and, predicated on a reduction in the sickness absence, you come to the conclusion that you need slightly fewer people, because your sickness absence is down?

John Keelty: Yes.

Q398 Mr Bacon: Right. Presumably, if you do not achieve your target, it is not in the end a question of money, because the people will have gone. The answer is an increase in backlog, isn’t it? It is a reduction in service.

John Keelty: It will have an impact on productivity, if that is to be the case.

Q399 Chair: Let me ask a question arising out of that. You intend to cut your staff by 19,000.

John Keelty: No, we intend to cut them over the spending review period by 10,000.

Q400 Chair: No-19,000, and then recruit 9,000. You have almost given away the first bit of the question. The 9,000 you are recruiting are all going to be redeployments out of the 19,000, are they?

John Keelty: Virtually all, yes.

Q401 Chair: How interesting. You are not taking the opportunity perhaps to change the quality or calibre of your staff; you are just going to move staff around the place in-house. We will take that as read. You are cutting the staff on personal taxation by 50%.

John Keelty: Not 50%.

Dave Hartnett: It is 34%.

Q402 Chair: Sorry-you are cutting by 34%. What have you done about modelling that in terms of a rise in non-compliance?

Dave Hartnett: We have done quite complicated modelling. Our top priority in relation to the reduction in staff in personal tax has been to minimise it in the first two years, so that we can complete the stabilisation of pay-as-you-earn, and to back-load. I will ask Stephen to come in, but our basic strategy is to continue to improve the performance of pay-as-you-earn to take savings from more effective processing through NPS. For example, last year, we were addressing eight tax years in one year, whereas normally for pay-as-you-earn we want to be addressing three: the end-of-year reconciliation for one year, the coding for the next year and in-year changes for another one. That will be a huge efficiency for us.

Q403 Chair: I understand that you think the workload will go down, because you will have dealt with the backlog. In effect, that is what you are telling us. But have you done an assessment for such a massive reduction in staff in the one section, and what it will mean for non-compliance? Have you modelled that?

Stephen Banyard: We do not believe it will have an effect-

Q404 Chair: At all?

Stephen Banyard: We do not believe so. Our first priority is to stabilise pay-as-you-earn. In the first two years, we are not planning to take many savings out of this area at all. In fact, we have put additional staff in to try to get the place up to date. The savings are back-loaded to years 3 and 4. They fall into a number of areas.

Q405 Chair: Mr Banyard, have you actually modelled it? I ask because one of the criticisms is that the NAO thinks you have a decent understanding of your costs, but in going forward you do not understand the relationship between expenditure and outcomes. I don’t quite know how you put it, but there is a way in which they put it in the report-

Jane Wheeler: The relationship between cost and value.

Q406 Chair: Thank you. I have taken this as an example of where you are going for a huge change-a 34% cut-because you think you will have dealt with the backlog and the system will have settled down. I really want to know whether you have modelled it. Or is this Committee going to come back to this in 2014, when we will still be the same Committee, and find that your non-compliance has gone up?

Dave Hartnett: We have done the modelling.1 We are also wary that there is a risk that voluntary compliance might drop.

Q407 Chair: By what?

Dave Hartnett: A few percentage points.

Q408 Chair: What? Can we see that? Can the NAO see the paper setting that out? What percentage?

Q409 Mr Bacon: A very small percentage could amount to a very great deal of money, couldn’t it?

Dave Hartnett: Indeed. Our objective is to maintain-

Q410 Chair: I can understand your objective. I can understand that. We all share that and we welcome it. But you are cutting your staff by 36%. You are going to have dealt with the backlog. We need to know what you think that means non-compliance will go up by. What percentage? What total? You have done the work. You ought to be able to share the papers with the NAO, which can then reflect that in its report.

Stephen Banyard: We have a new operating model for pay-as-you-earn operating under NPS. The system is new, and for the first time we have a whole-customer view. Our new operating model is therefore built around a completely new rendering of pay-as-you-earn. We are learning about that as we go along.

Q411 Chair: So you have not yet done the modelling to show the impact. That is slightly finger in the air, saying, "We think we’ll have dealt with the backlog, that the system will have settled down and that the data will be a bit better, so there will be fewer errors."

Stephen Banyard: No, not at all.

Chair: You have not modelled it. Mr Hartnett just said that you are expecting voluntary non-compliance to go up by a few percentage points.

Q412 Mrs McGuire: Could you explain what a full customer means rather than a half customer or a quarter customer?

Stephen Banyard: I can. Pay-as-you-earn was computerised in the mid-1980s, and it was computerised on the basis of employments. If you had an employment with someone, that employment information would be held on one of 12 regional databases. If you had two employments, the second employment could well be held on another database, and the only connection between them was pieces of paper.

Q413 Mrs McGuire: So you are looking at me now as an individual as opposed to an employee?

Stephen Banyard: Indeed. If you phoned our call centre-

Q414 Mrs McGuire: Assuming I could get through.

Stephen Banyard: Our staff would bring up your record for you and look at all your employments, as they do, and make the whole pay-as-you-earn tax right. That is a new facility for us.

Q415 Mrs McGuire: And that is working?

Stephen Banyard: Yes, it is working well.

Q416 Mr Bacon: May I ask if you have the facility to take on temporary staff when, for example, you have a sickness absence gap that was larger than you expected and therefore you have a backlog growing? Can you take on temporary staff to help sort out the problem?

Stephen Banyard: We can and we do, and we train them to a required level to do the work.

Q417 Mr Bacon: And where do you get them from? If they are temporary, they come and go-where do you get them from?

Stephen Banyard: We do not get them in the sense of temporary come-and-go; we would recruit people for a term appointment.

Q418 Mr Bacon: I see-on a fixed-term contract.

Stephen Banyard: On a fixed-term contract. We would bring them in for 11 months or 23 months, and we would give them full training. We have not had difficulty recruiting good people.

Q419 Mr Bacon: Why 11 or 23? So they cannot sue you for unfair dismissal?

Stephen Banyard: We have recruited for 23 months because that is the period we think we will need people for.

Q420 Stephen Barclay: How many people have you hired externally in the past 12 months?

Stephen Banyard: People or full-time equivalents?

Q421 Stephen Barclay: How many people have you hired externally?

Stephen Banyard: All of our staff have been recruited externally at some point.

Q422 Stephen Barclay: No, I said in the past 12 months. You are reducing head count to make staff savings. Over the past 12 months, how many people have you hired externally?

Stephen Banyard: In terms of people, rather than full-time equivalents, approximately 2,500.

Q423 Stephen Barclay: Right. I am conscious of the Cabinet Office freeze on external recruitment, but all those 2,000 people are defined as essential front-line staff?

Stephen Banyard: They are essential front-line staff, and they are helping us to get pay-as-you-earn up to date. We are determined to deliver a good system and a good service to the public, and that is how we are doing it.

Chair: As Stephen has just said, if you look at page 26, paragraph 2.20-

Q424 Stephen Barclay: Just coming back to that again, it is highlighted in the report, which states, at paragraph 2.20, "Wider experience indicates that organisations tend to over-estimate how much their cost reduction plans will actually save and best practice is to identify a contingency of 50 per cent." What level of contingency are you planning to identify?

Dave Hartnett: At the moment, certainly not 50%.

Chair: What?

Q425 Stephen Barclay: So you are not following best practice.

Dave Hartnett: We are not following what is here, because at the moment we cannot. I think, Mr Barclay, the key is we are monitoring risk incredibly carefully.

Q426 Chair: What is your contingency level? You said that you are now planning contingency, so what level are you up to?

Dave Hartnett: I do not have a level.

Q427 Chair: Do you have a number, Mr Keelty?

John Keelty: We are not planning for a specific number. As Dave says, we are monitoring the risk and we will adjust things as we get to them.

Chair: This is very waffly.

Q428 Mr Bacon: May I labour the point about the Rural Payments Agency? It introduced an enormous change programme at the same time as sacking all their most experienced staff and replacing them with temporary workers. You say that you are monitoring the risks, and it is good to hear that you are. What do you do about it when you identify that a risk that you have been monitoring has materialised? What happens next?

John Keelty: That is a very difficult question, because it could be any sort of risk. But as soon as that risk becomes an issue and we need to take action, we will certainly consider what is the appropriate action to take. It might be right to move another programme forward or, if the risk is a shortage of staff-we have 66,000 staff in the Department-we might have to move some staff from over there to here to help with that programme.

Q429 Stephen Barclay: Where you have identified savings, are you not moving as quickly as you can on those? From your answer to Mr Hancock’s question about property and to Mr Bacon, your contingency seems to be, "Ah well, we can speed up other programmes to deliver other savings."

Mr Bacon: Move things forward.

Stephen Barclay: Surely you would be delivering those as quickly as possible anyway.

John Keelty: We have a history of handling large programmes-we are a large Department handling large change. Throughout the past few years, as necessary, when things ebb and flow, we adjust our programmes accordingly. I am having difficulty answering your question, because it is a largely hypothetical situation.

Chair: Okay. Maybe the Comptroller and Auditor General can put it in another way.

Amyas Morse: I will try to. Everything you’ve just said is fair-you have a large Department and you have a record of handling large programmes. It is also fair to say that you have been able to take a reasonable amount of time over doing that. That is not a criticism; it’s just a factual statement and, I think, a fair comment. You now move into an era in which you are trying to change a lot faster than you were. I think that is fair to say.

Dave Hartnett: It is.

Amyas Morse: And you have been able to carry out what I call fairly discrete change programmes, but now the connection between one another is becoming much closer. So what we are trying to say-I urge you to consider this and I shall be interested to hear your reaction-is that you must be able to react early, to understand the linkages early and to get whatever actions you are going to take into place in time to retrieve the situation before the performance gap is too great. It is not that we do not think you are trying to tackle those things, but you could do with developing a more sophisticated and integrated approach as early as possible.

John Keelty: We do recognise the need to look at the risk and take an integrated approach. You have recognised in your Report that we are setting up the governance arrangements to do that. The question is: can we set aside a chunk of money for the eventuality that something might happen? As a Government Department in times of austerity, finding a chunk of money that you can just set aside is extremely difficult. The Government want to use that money in other ways.

Amyas Morse: I have every sympathy with what you have said. A chunk of money not doing anything should be as small as possible. But that is a result of how fast you can react-how quickly you can spot if you are off plan. We are talking about what you would do. How big a gap is there from flash to bang? If something is off track, are you geared up to react really quickly and put those extra measures into place, or will quite a long time gap be allowed?

Dave Hartnett: We are geared up to identify any suggestion that a project is starting to go that way. With regular, weekly reporting, we have put change directors into each of our major business areas, and that monitoring process is working very effectively.

Q430 Mrs McGuire: But it is not about identifying the risk or the problems; it is about how quickly you can react. Is HMRC now flexible enough to give a far quicker response when a risk is identifiable? Have you learned any lessons, dare I say it, from the child tax credit fiasco, where the risks were obvious to everyone and HMRC was so far behind the curve that people are still recovering from the damage done?

I wonder what business people listening to this exchange would think of it, because HMRC demands precise information, yet there is a lot of woolliness in our discussion this afternoon. We have heard, "Contingencies are not there. We are not sure what we will do if this happens. We are monitoring things." Frankly, you guys are in the front line not just of tax, but of tax and benefit, child benefit-the lot. People’s financial viability will depend on your being flexible enough. Frankly, I do not think that you have given us much confidence today that you do have that flexibility. Maybe I am alone in that.

Amyas Morse: May I just make one comment? I was listening to a presentation recently by your change director, Phil Pavitt-is that right?

Dave Hartnett: Yes. Chief information officer and change director.

Amyas Morse: I have to say that if the performance matches everything that he was saying-I am sure that it does-then it sounded pretty impressive to me.

Chair: He is not giving evidence to us Amyas

Mrs McGuire: There is always a gap between rhetoric and delivery. We have three senior people here from HMRC and I am just not getting enough from them, which makes me feel that, within the next two or three years, they will be implementing major changes in terms of the flexibility of response.

Q431 Chair: I think Anne has summed up the view of the Committee very precisely.

I want to ask a final thing on tax credits. My understanding is that the current debt is £4.7 billion. I have two questions. How much of that is due to customer error, and how much is down to HMRC error?

Dave Hartnett: Most of it is due to customer error or worse.

Q432 Chair: What proportion?

Dave Hartnett: I will have to write to you, because I cannot remember the number.

Q433 Chair: In a week, please.

Dave Hartnett: Of course.

Q434 Chair: Can you give us a ballpark figure?

Dave Hartnett: I will give you a figure, but I may be wrong: 80%.

Q435 Chair: So that is customer error. How much is realisable out of that £4.7 billion?

Dave Hartnett: We think probably something of the order of three to three and a half currently, but we are working on that. We have taken a new approach to debt generally, and we are aiming to reduce the debt to £3.7 billion by the end of the spending review.

Q436 Chair: It is going up at the moment. You will reduce the debt by getting the calculation right in the first place with this cohort of people, not by pursuing the debt. My understanding of paragraph 38 on page 11 is that you have only collected £380 million against your target of £550 million. So you are 50% down on your target. Then there is this funny little sentence in the Report-I did not mean peculiar-that there is £1.7 billion of outstanding debt that you are remitting.

Dave Hartnett: We are certainly planning to remit a substantial number.

Q437 Chair: £1.7 billion?

Dave Hartnett: I am trying to find it in the Report.

Currently, we are remitting £1.055 billion relating to tax credit customers and that is in relation to debts incurred between 2003-04 and 2008-09, where we have received no payment in the past 12 months, and some of those people we cannot find.

Q438 Chair: So we have lost over a billion, and we might lose another £700 million or whatever it is.

Dave Hartnett: I don’t know yet.

Q439 Chair: When are you going to take that decision?

Dave Hartnett: Whether to remit the £1.055 billion?

Q440 Chair: No. There is a figure in the Report. I think you have been pointed to it. What page is it? Where did I pick this up from?

Amyas Morse: Page 47.

Chair: Page 47. It’s in the other Report. "The Department estimates that £1.7 billion of new tax credits debt is likely to be generated"-blah, blah, blah. Is that where I got the figure?

Dave Hartnett: Yes, but that is not about remission.

John Thorpe: That is not about remission. Based on the Department’s own calculations, "without any further intervention" or action to reduce debt, "debts could increase to £7.4 billion by 2014-15."

Dave Hartnett: I am terribly sorry. I have forgotten the question.

Q441 Chair: It’s a heck of a lot of money is what we are saying. My own view, which I am sure is shared by the Committee, is that the only way you really get this right is by getting the calculation right in the first place, not by trying to pursue really poor people.

Dave Hartnett: I agree, and our strategy is to help our customers and ensure that our calculations are right in the first place.

Q442 Chair: But you are remitting over a billion?

Dave Hartnett: That is the plan.

Q443 Ian Swales: I want to ask a couple more things about tax credits. Just to put this into context, what is the entire operating cost of HMRC, just to the nearest billion?

John Keelty: £3.5 billion.

Q444 Ian Swales: So about £3 billion of staff costs. Would that be about right?

John Keelty: A little less than that.

Q445 Ian Swales: I think it is really important to have that number in context when we are talking about the size of these numbers.

John Keelty: £2.2 billion.

Ian Swales: £2.2 billion is the staff cost. Your entire staff costs are £2.2 billion-

Q446 Chair: And the rest takes you up to? That is huge. What was your original figure to Ian?

John Keelty: £3.5 billion, so on IT and accommodation-

Ian Swales: £3.5 billion, so that is all the other costs of offices and so on.

Chair: A third and two thirds, really.

Q447 Ian Swales: Okay. I just want to pull out a few figures from this. It says in paragraph 33, page 10, that you have a target to reduce tax credits error and fraud to no more than 5%. My calculation is that that is about £1.4 billion-that is your target to get down to, for errors and frauds. Is that a fair assumption? I am taking 5% of the £28 billion spent on tax credits in the previous paragraph.

Dave Hartnett: We actually have a higher target each year, Mr Swales, because the number of new claims coming into the tax credit system is about 20% a year. So our target is to bring this down by £1.4 billion by 2014-15, but we will do that by tackling error and fraud at roughly £2 billion a year. We have changed the whole strategy to do that. Whereas you heard Stephen Banyard talk earlier on about "process now, check later", the big change in tax credits is to check now, process afterwards.

Q448 Ian Swales: It says: "The Department has a target to reduce Tax Credits error and fraud to no more than 5%", so what are we talking about? Some £1.5 billion or so as your target-is that right, or have I got the wrong number? What would your target be?

Dave Hartnett: £1.4 billion.

Q449 Ian Swales: So it is £1.4 billion-that was the first figure I said. Okay. So that is one figure that is important. Let us then move to the one that the Chair just mentioned, which is that you are going to assess-in paragraph 38, on the facing page-"the value for money of collecting £1.7 billion of tax credits debt not under active recovery". Can you just talk us through how you are going to assess the value for money of collecting that debt?

Dave Hartnett: Broadly, we assess the value for money by looking at what we think we are going to recover as against the cost of the actual recovery. That is a balance that we are making all the time in the area of tax credit.

Q450 Ian Swales: So when you say "the cost of making the recovery", what kind of things are we talking about? People and resources, presumably.

Dave Hartnett: People and resources. We have outsourced some debt collection to debt collection agencies.

Q451 Mr Bacon: Can you just remind us how much? On the record.

Dave Hartnett: I cannot-

Q452 Mr Bacon: It was £214 million, wasn’t it? That you have outsourced.

Dave Hartnett: I think it was more across our whole business.

Stephen Barclay: Well, it was £214 million between 27 July 2010 and 14 January 2011-

Q453 Mr Bacon: So that is in the last full year. And, of that, 26% or £57 million was recovered. Well, if you are adding an extra level of new tax debt of £1.5 billion in a year, and you are getting £57 million through recovery, it doesn’t take Einstein to work out how those two graphs are going to diverge, does it?

Ian Swales: In fact, the Report says that "without any further intervention tax credits debts could increase to £7.4 billion by 2014-15." That is in paragraph 37. So what intervention are you going to make?

Dave Hartnett: We have changed our approach to managing tax credits. The four key things are: first, checking first, paying second; secondly, reviewing the risky areas of tax credit, such as the cost of child care, which is one of our highest risk areas; third, by having a process of cleansing, which is automatic checking of claims when they are made, for which we have a new risk tool; and, finally, ensuring that we improve the training of our people and how they work with claimants, so that there is less over-claiming.

Q454 Ian Swales: Okay. Much nearer the end of the Report, on page 46, in the detailed tax credit area, it discusses your having launched a joint fraud and error strategy with the Department for Work and Pensions, targeting £8 billion of tax credit losses over the next four years. How much resource, in people and money, has HMRC put into that?

Dave Hartnett: I do not know the answer to that.

Q455 Ian Swales: Well, approximately. Is it 10 people? Is it 100? Is it 1,000?

Dave Hartnett: No, it will be hundreds.

Q456 Chair: Does Mr Keelty know?

John Keelty: No, I don’t know the exact number.

Chair: Does anyone behind you know?

Q457 Ian Swales: We are talking about something that is about £8 billion of your tax credit losses. I would have thought that you would know something about this project that you have set up. Do you have any metrics at all on it?

Dave Hartnett: Not today, I am afraid. I am very sorry.

Chair: Very finally-

Dave Hartnett: The information has come from behind. We have 1,700 people or full-time equivalents in tax credits working on compliance activities.

Chair: That is on tax credit.

Q458 Ian Swales: Is that on this joint project with the Department for Work and Pensions?

Dave Hartnett: No, we do not have that number with us.

Q459 Ian Swales: Could you write to us?

Dave Hartnett: Of course.

Q460 Ian Swales: I made the context-setting remark at the start that the entire manpower of HMRC cost only £2.2 billion. My final question is: if you had your time again, would you have cut your staff in HMRC as fast as you did? Do you need to revisit your future staffing strategy in the light of all the issues that we have discussed today?

Dave Hartnett: There is a real balance there. If you have look at what we have achieved, despite cutting, we cut our compliance resource by 20%. In five years, while that was going on, we doubled the compliance yield from £6.9 billion or £7 billion to £13.9 billion. Stephen talked about how we had improved the post. Our contact centres are nothing like good enough yet, but there has been a huge improvement there too. We have taken our debt balance down by £3 billion to £4 billion. We are doing good things.

Q461 Ian Swales: Other than the well publicised £900 million, you do not think that there are any more "invest to save" projects for HMRC, given the tax gap that we have talked about before.

Dave Hartnett: No, Mr Swales. We are constantly looking for projects. Whether we can get the funding for them is another matter. Some of our offshore work-in the last week we have made it public how we are approaching the 6,000 account holders in a particular Swiss bank-we will be looking to fund.

Q462 Ian Swales: Thinking about some of the things that we have spoken about this afternoon, have you raised any "invest to save" projects with the Treasury that have been turned down?

Dave Hartnett: Not that I can recall. The reason for that is that we want to demonstrate clearly what we can do with the £917 million first. We are ahead of profile in what we promised to deliver.

Q463 Ian Swales: But the £900 million, as I understand it, is specifically to do with tax avoidance. We have many other issues, such as recovery of debts.

Dave Hartnett: No, there are lots. It’s debt recovery, avoidance, and principally evasion.

Q464 Stephen Barclay: Quickly to pick up on a question, it is very misleading to talk about the compliance yield in isolation. What Mr Swales was driving at was the cost of staff compared to the amount of tax missed that could have been collected. If a company owes £5 billion in tax and you do a handshake deal with them for £1 billion, the compliance yield looks exceedingly good-we have brought in £1 billion. What it does not show is that we have lost £4 billion. So talking just about the compliance yield does not address the point that Mr Swales was driving at, does it?

Dave Hartnett: I think, Mr Barclay, that you know that I am going to say this: we don’t do handshake deals. I cannot think of a case like the one you have just described.

Q465 Stephen Barclay: I was just quoting internal notes. Whether it is done on a handshake or done through governance, the point stands. In talking about the compliance yield in terms of the optimum level of staff as against the amount of tax you collect, we also need to see how much tax could have been collected but wasn’t collected. That’s correct, isn’t it?

Dave Hartnett: I think that that’s a very fair point.

Chair: Matt has a question on the compliance issue.

Q466 Matthew Hancock: No, it’s on the broader point that Mr Swales raised. Something bugged me during the discussion on the savings that you are making and the contingency, which is linked to this point. Could you set out clearly where you are on the milestones to achieving that? The contingency is an important question, but I don’t think we got you on the record in that discussion with where you are up to in delivering the goals that you set out.

John Keelty: Let me try to answer the question. Please tell me if I am not answering it.

Matthew Hancock: Was I not clear? Are you on track?

John Keelty: We believe that we are on track, yes.

Q467 Matthew Hancock: You believe that you are on track, or you are on track?

John Keelty: The mid-term review that we had of the change programme showed that we were on track. Like all big programmes, we need to take action to make sure that we keep going on the right track. That Report said that we didn’t need to make any major changes to the direction we were heading in. So we are on track. With the large portfolios of change, there will be things that are doing a little better than others, but we are on track.

Chair: That is it. You end on a positive. Thank you very much for coming, and I am afraid to say that I think that there will probably be some more in the near future.

[1] Real Time Information will improve voluntary PAYE compliance because reporting to HMRC will be integrated into the payroll process and employers will know that HMRC has earlier visibility of payments due

Prepared 19th December 2011