Treasury Committee
Oral evidence: Tax credits, fraud and error, HC 795
Thursday 27 October 2016
Ordered by the House of Commons to be published on 27 October 2016.
Members present: Mr Andrew Tyrie (Chair); Mr Steve Baker; Helen Goodman; Stephen Hammond; George Kerevan; Jacob Rees‑Mogg; Wes Streeting
Questions 1 - 113
Witnesses
I. Jon Thompson, Chief Executive and Permanent Secretary, HMRC, Nick Lodge, Director General, Transformation, HMRC
Jon Thompson, Chief Executive and Permanent Secretary, HMRC, Nick Lodge, Director General, Transformation, HMRC.
Q1 Chair: Thank you very much for coming to give evidence to us this morning. I was going to start by looking at making tax credits digital, but I should start by asking you for an update on the correspondence that I have had with you about this subject. Your letter of 21 October said that as of a couple of days earlier you had concluded around 17,000 of the 24,000 requests for review of Concentrix’s decisions and implied that none of them were more than 10 days old. Was that right?
Jon Thompson: That was the situation then. That is correct, yes.
Q2 Chair: What is the situation now? We did ask for weekly updates. Where are we at the moment?
Jon Thompson: The situation now is that we have received approximately 26,000 requests for mandatory reconsideration. We have closed just over 20,000 of those. We are currently working on those received on 15 October, so that is well within the number of working days target. We are about seven or eight working days behind today. The target is to do them in 21 days, so we are well within that target.
Q3 Chair: What is your best estimate of the number of errors that have been made?
Jon Thompson: Do you mean the number that are being overturned in favour of the customer?
Chair: Yes.
Jon Thompson: On the high-risk renewals for this year, it is just over 90% in favour of the customer.
Q4 Chair: It is pretty shocking, is it not?
Jon Thompson: Yes.
Q5 Chair: It is hardly surprising people are very upset.
Jon Thompson: No, it is hardly surprising. I gather that you would know that I have given evidence to both the Work and Pensions Committee and the Public Accounts Committee on this. In fact, we came to the Public Accounts Committee only yesterday. Let me repeat the situation. First of all, on behalf of HMRC I would like to apologise to customers, because clearly this is a failure of basic customer service. If you are unable to answer the telephone at the required speed, it is unsurprising that benefit claimants can get upset about the ultimate result of that. I am very happy to go through why this occurred, how it occurred, the date they occurred on and so on in the same way as I did in both the Work and Pensions Committee and PAC yesterday.
Q6 Chair: Colleagues may take us into that territory or they may not. I have a couple more questions. What estimate have you made of fraud that may have crept in in addition to error?
Jon Thompson: Do you mean in general?
Chair: With respect to the opportunities that have been created for fraud by the fact that this has clearly been very badly handled by Concentrix.
Jon Thompson: I am slightly struggling with the question. If you step back, the point of this contract was to provide HMRC with additional capacity and capability to reduce the overall fraud and error rate in the tax credits system. If you go back three or four years, the overall rate was between 8% and 10%. Ministers thought that that was too high and targeted it being below 5%. In order to do that we needed additional capacity and capability to basically do more work in this area to reduce the rate.
That was the original motivation for bringing in a private sector partner to add more people with different analytical skills. The overall fraud and error rate is now down at 4.8%. As I said to the Work and Pensions Committee: if you were the Chancellor of the Exchequer, for the investment of just over £27 million in this contract, the return to the Exchequer is some £270 million. Those are both estimated numbers.
Q7 Chair: To what extent will making tax credits digital help prevent fraud and error?
Jon Thompson: It should help, because around 70% of the challenge here on error and fraud is because of incorrect reporting or failure to report changes in circumstances. We have introduced an online service so that you can report your change in circumstances to us, because one of the design faults of the tax credit system is having an annual award is quite a long cycle. If your life changes, your circumstances change, your income changes or your child costs change, then your award changes. Leaving it a year is too long, and so we have introduced an online “report your change in circumstances”, so we can correct your award in-year and you can correct your award in-year. There are definitely some advantages in digital working in that we can keep your award up to speed with your life.
Chair: Okay. Those are helpful, relatively brief answers on which others may invite you to expand.
Q8 Stephen Hammond: Good morning. On the Revenue Benefits website, under tax credits, it says, “The definition of a couple is set out in legislation.” In your own letter to the Chair of the Public Accounts Committee in July you described it as an inexact science. Given the complexity sometimes of applying those rules in terms of whether people are living together as husband and wife and the sensitivity of the matter in terms of the undisclosed partner investigations, did the Revenue give guidance to Concentrix on best practice? Do you think they followed that best practice? In hindsight, if they did not follow best practice, where did they really fall down?
Jon Thompson: Do we give them guidance on best practice? Yes.
Q9 Stephen Hammond: Were there clear best practice guidelines set out for them to follow?
Jon Thompson: Yes. I am sure Nick can go through the detail, if you want to. The plain answer to your question is yes.
Nick Lodge: Shall I just add a little bit?
Stephen Hammond: It would be helpful to have some colour. Mr Thompson’s answers are helpful considering some of the witnesses we have who go on rather longer.
Nick Lodge: Concentrix were given access right at the start of the contract to our technical manuals and guidance. We help them with all of the training. We were there on an ongoing basis to help with what is, as you rightly say, a very complicated area. We called them floor walkers; we had people in the Concentrix premises for quite a few months through the early part of the contract to help their staff understand how to address these kinds of issues. There was quite a lot of help and support and certainly access to the technical guidance and to best practice.
Q10 Stephen Hammond: You are clear that out the outset they were given every opportunity to follow best practice.
Nick Lodge: Yes.
Q11 Stephen Hammond: You helped them with that. At what stage did you do a review as to whether best practice was actually being followed in those undisclosed partner investigations?
Nick Lodge: As part of the regular monitoring under the contract, we would look at a sizeable sample of claims and decisions on those claims each and every month, and we would assess that and feed back, working with Concentrix, to their staff and managers if anything was not being done to the standards that we require. There is a very regular monthly monitoring and reporting of the quality of decision-making.
Q12 Stephen Hammond: What was in their contract that you could enforce to bring them up to that best practice? I am taking it from your answer you were monitoring and you recognised that there were failures or ongoing failures to meet best practice.
Nick Lodge: Quality standards varied a bit month to month, but for the most part since October 2015 we have been in the 90s in percentage terms. There are contractual terms that affect the amount that Concentrix get paid if their quality falls below 97%. That was why we were doing monitoring regularly. We were looking at 500 cases most months, but not every month—it depended on the nature of the work. But certainly there was regular monitoring and regular weekly meetings to talk about performance, including those aspects.
Q13 Stephen Hammond: You accept that best practice in the end failed.
Jon Thompson: Yes.
Q14 Stephen Hammond: You are clear now where it fell short and about the remedies that need to be put in place in the next contract.
Jon Thompson: Can we separate those two questions?
Stephen Hammond: We can.
Jon Thompson: Let me be consistent, because this is the third time I have given evidence to a Select Committee on this. The basic problem here is that you have a process whereby somebody makes a claim. There is, we believe, some third-party evidence that would suggest that we ought to make an enquiry about whether that claim was accurate or not. You write to the benefit claimant. You ask for further information. It is perfectly natural that somebody’s reaction therefore is to think, “I will ring.” What happened here is they were not able to get through on the phone.
To be consistent with the evidence I have given others, on the week commencing Monday 15 August the speed of answering the telephone reduced dramatically. A large number of enquiries go out. People naturally ring. They cannot get through. If you know anything about queuing theory and how contact centres work, it is perfectly natural that, if you cannot get through, you ring again. There have been some stories about people ringing 60 or 70 times and spending 45 minutes on the telephone.
That is therefore an element of the tax credit system that is fundamentally failing. You cannot resolve your case if you cannot speak to somebody to resolve it. For the last two weeks of August, essentially the speed of answering the telephone within five minutes was under 10%. That is the critical period after the renewal of your tax credit claims finishes in July—that cycle—and therefore you get into this situation where benefit claimants are being asked to justify their situation. They are trying to get through. They cannot get through. You enter into a spiral where you cannot get through multiple times. That then leads to the cases that are being reported. I have here what we then did when we stepped in, how that worked and so on and so forth, which I am happy to go through.
Q15 Stephen Hammond: One of the other things that happens when people cannot get through on the telephone is they write to people like me or my colleagues. One of the things we are keen to understand, because I have pursued a number of these cases on behalf of constituents, is the exact burden of proof that is then required. You are quite right: if someone comes under investigation, it is only right that you investigate them. Is the burden of proof clearly set out and are you clear that Concentrix was following that burden-of-proof test correctly?
Jon Thompson: There are two burdens of proof: one when you make your claim, and then separately if you want to change your claim. Under the Tax Credits Act 2002 the claimant needs to provide a declaration to show that they are entitled to receive tax credits. The original onus is upon the claimant to provide all the necessary information for us to calculate the correct amount of award. If we then want to change that, the onus switches to us. Under Section 16 we must reach a threshold, which is to have reasonable grounds for believing that the information you are giving me is not correct. That is the test. There are two different tests. There is your original claim and then there is if we want to change it.
Q16 Stephen Hammond: “Reasonable grounds” is again an inexact word or inexact phrase. Are you clear that Concentrix knew exactly what reasonable grounds meant? Are you clear that they were undertaking that test in the correct manner?
Nick Lodge: That is the sort of thing that would have been included in the information that we had shared with them in the training, the technical manuals and the materials. The answer to the first question is yes. Our quality checking that I described earlier did look at a significant number of decisions each month, and by and large Concentrix were getting those decisions right—not every single one, but they by and large were getting those decisions correct. They were following the processes that were laid down under both Section 16 of the Tax Credits Act and Section 18 of the Tax Credits Act that Jon mentioned.
Q17 Mr Baker: Mr Thompson, you just articulated with admirable clarity what it is like for people to be in the queuing system. I really appreciated that. Mr Hammond just indicated that people do get in touch with us. You do appreciate, do you not, that people of modest means stuck in these cycles can be reduced to states of really profound anxiety and despair?
Jon Thompson: Yes.
Q18 Mr Baker: What can be done, and how quickly, to end this problem? What exactly is it that you are going to do? How energised are you to end this human cost of these systems failing in this way?
Jon Thompson: Let us be clear about what we have already done. As I said, this issue essentially started in the second week of August. I was alerted on Monday 5 September. There is a lesson in there about it taking three weeks, essentially, to get to me. Once we stepped in on the week commencing 5 September, we deployed extra HMRC staff. We first of all deployed 50 additional HMRC staff to go through the mandatory reconsiderations, which the Chairman was asking about. We then added another 100 staff into that and then we added a further 200 staff in order to improve services on the telephone line. We are trying to fix the problem, which is people trying to get through on the phones, and we are also trying to speed up the mandatory reconsiderations process, because it is that process that essentially gets you back into payment.
As soon as we became aware, we stepped in and we think we did something. We took all new work away from Concentrix and 181,000 incomplete cases were taken back by HMRC. We have managed to complete 178,000 of those cases. The vast majority of this is now being done by HMRC. We are not still in the situation we were in.
Q19 Mr Baker: Okay. I do not want to trespass too much on my colleagues’ questions, but it just occurs to me that you are clearly aware of the human cost of these systems breaking down, and I am grateful to you both for nodding. Perhaps we will come back to it later. Could you just remind us, Mr Thompson, of the constitutional position of HMRC?
Jon Thompson: “It is complicated” is the best way of putting it.
Mr Baker: Briefly perhaps.
Jon Thompson: It is a non‑ministerial department where the commissioners have certain powers as defined by the 2005 Act, which created Her Majesty’s Revenue and Customs. In addition to that, under the Government Resources and Accounts Act I am the accounting officer for all aspects of public expenditure. As it was put to me by the PAC yesterday, in the end the buck stops with me as the Chief Executive and first Permanent Secretary.
Q20 Mr Baker: Some decisions are yours and yours alone to take. Is that right?
Jon Thompson: Yes. Unlike a department with Ministers, there are some decisions that are mine and mine alone. There are some that are for the commissioners and there are some that require Treasury Ministers’ approval.
Q21 Mr Baker: What caused you to decide not to extend the Concentrix contract and to notify them at very short notice?
Jon Thompson: It seemed to me that this situation was frankly untenable. We are badly letting down customers. This is a very customer‑centric, very human aspect of public services. There is basically in my opinion a breakdown of trust, and I stopped all further conversation about extending this contract. Yesterday I informed the PAC, and I believe the Financial Secretary told the House, that we are in late-stage legal negotiations about a mutual termination of this contract.
Q22 Mr Baker: I was about to ask you, “Why not terminate it?” You have given us the beginnings of an answer there: that you are looking for legal reasons for a mutual termination rather than an imposed one. Is that what I am to understand?
Jon Thompson: It is in my opinion generally better to have a mutual termination, because it is in the customers’ interests and in the taxpayers’ interests to have a negotiated exit, because it means that we can get all of the necessary information, case information and some of the knowledge that exists. If you go into a legal dispute, you can put some barriers in that you can remove by having a negotiated mutual termination.
Q23 Mr Baker: Given what you have said about the extent to which you have taken cases back in-house and so on, Mr Lodge, what is left for Concentrix to do?
Nick Lodge: Hardly anything. They have just about cleared the relatively few cases that we left with them. Originally that was 15,000 cases. If they had not cleared them by close last night, it is a matter of a day or two and then they will no longer be doing any tax credits work on claimants’ cases.
Q24 Mr Baker: Have either of you formed a view on whether or not HMRC should continue to be a non‑ministerial department or should Ministers be more involved in decisions of this political interest?
Jon Thompson: I am happy to work under any constitution that ministers decide. I have worked for two non‑ministerial departments now—Ofsted and HMRC—and I have worked for two ministerial departments. I am reasonably fluid about it. Ministers decided on this. There are other systems that operate elsewhere in the world. I will happily work in this system until Ministers tell me otherwise.
Mr Baker: Thank you very much.
Q25 Helen Goodman: Two of the constituents who contacted me with their Concentrix problems were victims of domestic violence. Does the Revenue for its other work have a protocol for how to deal with domestic violence victims?
Nick Lodge: We take that very seriously. If we know that somebody is a victim of domestic violence—if they have told us about that, and this extends to tax credits as well, certainly, in terms of the records HMRC holds—then we will make sure that we make it clear on that record that that is the case, so that we can act accordingly in what can often be very difficult circumstances. The very last thing we would ever want to do is to put somebody at risk.
Q26 Helen Goodman: Following up on the point that Stephen Hammond made about best practice, did you share with Concentrix guidance on how to deal with domestic violence cases?
Nick Lodge: We certainly shared with Concentrix guidance on how to deal with the most vulnerable people that they might deal with. I would have to absolutely double-check on the specific domestic violence points, but we certainly did share with them guidance on how to deal with the most vulnerable—how to deal with people who might contact them who would be extremely upset, for example.
Q27 Helen Goodman: One of the pieces of evidence that was given to the DWP Select Committee was that you did not note as high risk single claimants who had been single for 12 months, but people fleeing violent partners will be doing that in the here and now, so I would be very grateful if you could do that.
Another group of people who have particularly suffered because of Concentrix’s maladministration are people whose children have disabilities. In one case I had a constituent whose child had a lactose intolerance, so when I suggested that they go to the food bank they said, “That is not a good idea, because then we cannot control what our child is eating.” In another case a child had a congenital heart problem and had had seven open-heart operations. Their credits were suspended for 12 months. Given that the rates for tax credits are different where children have disabilities, there cannot really be any excuse whatsoever, can there, for Concentrix not seeing that these people were vulnerable?
Nick Lodge: The records that they would have had might have revealed that through the rates that were being awarded.
Q28 Helen Goodman: Surely it would. You see what the award is every week. You can tell from that, can you not? Or are you saying that they might not have known?
Nick Lodge: They might not have been able to see that easily, but certainly people should not have been without money for the length of time that you mentioned there. I hope those two particular cases that you mentioned have been resolved now, but if they have not we would certainly take the details and have a look at them.
Q29 Helen Goodman: Are you saying that when you passed over the information you did not flag where there were either adults or children with disabilities in families?
Nick Lodge: I would have to double-check, but I think the data that Concentrix would have received would not necessarily have flagged that level of detail. I will double-check that for you and let you know if that is wrong.
Q30 Helen Goodman: In the hearing in front of the DWP Select Committee Concentrix said that you handed over to them a list of 1.5 million people that the Revenue described as high risk. I would like to ask you how it could possibly be the case that 1.5 million people were high risk given there are only 4.4 million people getting tax credits. You are surely not suggesting that a third of recipients of tax credits are people who are potential fraudsters?
Nick Lodge: Not potential fraudsters, but in these cases we took the very detailed statistically valid annual sample that we do every year that tells us what is going on in the tax credits system, and explains where the error and fraud is. That is published each year. Our analysts can therefore look at the characteristics that are typically associated with claims that are incorrect for one or another reason. I should just say that much of it is error, because of the complexity of the tax system, and not necessarily fraud.
What we are doing with those 1.5 million cases is effectively saying, “These are cases that share the characteristics of cases that we know often are incorrect.” We are not saying that they are all incorrect by any means, because clearly they are not all incorrect, but these are the types of cases where people might have made mistakes or there might be incorrectness in them.
It was from those cases that Concentrix would then apply their much more detailed analytics using credit reference agency data and so on and so forth to select those cases that they thought we were worth checking up on. That would have been a much smaller number than the 1.5 million, because you are absolutely right: not all of the 1.5 million will have been incorrect. Concentrix would have then taken from that 1.5 million cases those cases that on their further analysis they thought were worth checking.
Q31 Helen Goodman: Yes, I agree with you. They whittled it down to 324,000. You had a cost-benefit analysis of how good this piece of work was going to be for the taxpayer. Your assumption was that there were going to be savings of £400 million, which there have not been, and you described a third of tax credit recipients as high risk. Throughout the tax and benefits system the rate of fraud and error is about 3%, so how could you possibly have imagined that it was going to be 10 times as much in this particular area? Why on Earth would that be the case for this group of people as opposed to everybody else who claims benefits?
Nick Lodge: The rate of error and fraud in the tax credit system, as Jon said, on the last figure for 2014-15 was 4.8%. That is the value. That is the £1.37 billion cost of error and fraud in the tax credit system. The number of individual awards that are incorrect is a slightly higher figure than that or a rather higher figure than that—probably about 20% are incorrect at some point in the year. That is why we check up a large number of those cases.
In flagging the 1.5 million cases as at risk, we are not saying that they are all wrong—definitely not. We are saying that they might be wrong and they might need to be checked. As I said, often it is not fraud. It is an innocent mistake by claimants having to deal with what is quite a complex system. We are looking at the characteristics and profiling and using our analysis to say, “These are the kinds of cases that are at risk of being wrong.”
Q32 Helen Goodman: The discussion with the DWP Select Committee focused on these 324,000 people to whom Concentrix wrote, but in answer to a parliamentary question from me, the Minister said that 381,000 letters had been sent on single status, 254,000 had been sent on hours worked and 312,000 had been sent on whether childcare was accurate. That is a total of 947,000, which is nearly 1 million. It is three times as many as what has been discussed in Parliament up to now. Could you tell us what is happening? We are now beginning to get through the people that have had the hours-worked letters in our constituency office. What is happening with these other 600,000 people?
Nick Lodge: We have two sets of figures that are both correct but are measuring what is happening over a different period of time. The 324,000 would have been what was happening since April 2016 through to August and September this year. The figures that you just mentioned—obviously I do not have them in front of me—sound very much to me as though they are figures since the beginning of the work that Concentrix did from November 2014. I think they are both correct but for different periods of time. As I said a little while ago, Concentrix no longer have any of these cases, or if they do it is literally a handful live on their systems, and they are now all through and completed.
Q33 Helen Goodman: Given that you have not got the numbers, it would be very helpful if you could send a note to the Committee about the outcomes, the proportion of mandatory reconsiderations, the proportion that were right and the proportion that were wrong for the hours-worked people and for the childcare challenges as well, because we do not want to see this repeated twice further. Mr Thompson, in the light of this experience, do you think there might possibly be something unethical about having a payment-by-results contract for removing people’s tax credits?
Jon Thompson: One of the five lessons that needs to be learnt from this is whether an incentives-based contract was the right way of going about this, yes.
Q34 Helen Goodman: Would you do it again with another aspect of HMRC activity?
Jon Thompson: I would not do it again on this. Mr Hammond was inferring that we would let another contract. We have no intention of letting another contract.
Q35 Helen Goodman: Good. I am glad to hear that. You are improving the administration and people are getting back payments. But when they are getting their back payments, many people are not getting the full back payment to which they are entitled. For example, their credits were wrongly stopped for, let’s say, eight weeks. Let’s say they lost £1,500, but when the cheque comes they get £500 followed by £300, followed by you know—instead of getting it in a lump sum. What has the cash flow benefit been to HMRC by not paying back what people are owed straight away as soon as you find out about it?
Jon Thompson: We have not worked that out. We gave a commitment to Work and Pensions to look at the way in which the money that is in that gap period, in your eight weeks, goes back to claimants. We gave a commitment that we would look at that, but we have not looked at what the cash-flow benefit is.
Q36 Helen Goodman: There obviously is a cash-flow benefit to you if people are owed money, and instead of giving it to them as soon as you discover it, you are dribbling it out to them over three months or something, is there not? There obviously is a cash-flow benefit.
Jon Thompson: Yes. We have not calculated that. My focus and our focus has absolutely been on sorting out 181,000 cases. I have not really considered whether there was a cash-flow benefit for us. My focus was on sorting out the situation for as many claimants as possible as fast as possible.
Q37 Helen Goodman: Good. I am glad to hear it. You should find out, because it will leave many people with the lingering suspicion that HMRC was not as alert to tackling Concentrix’s underperformance because HMRC had some skin in the game as well.
Jon Thompson: I would absolutely refute that whatsoever. We did not go into this thinking, “What would be a really good idea is to have a complete crisis of confidence about the tax credit system because there may be some marginal cash-flow gain for the Government.” I reject that.
Q38 Helen Goodman: No, but you did go into it telling Ministers that there were going to be £400 million worth of savings, did you not?
Jon Thompson: Yes, that was the estimate.
Q39 Helen Goodman: How did you make that estimate?
Jon Thompson: Your suggestion is that somehow we did this in some way that would give the Government some marginal cash gain—I completely reject that idea. Sorry to be so clear about it. I reject it.
Q40 Helen Goodman: Let’s find out what the numbers are and we will see whether it is a marginal number or a significant number.
Jon Thompson: I am happy to give you the number. I am sorry to keep pushing you back. It is the suggestion that we somehow deliberately did this. I completely reject that idea. There may well be some marginal cash gain and we will try to work it out for you, but to continue the inference that we did this deliberately because there would be some marginal cash gain for the Government, I am sorry, I have to reject that idea.
Q41 Helen Goodman: No, you were trying to make savings by reducing fraud and error, and Concentrix went about it in an extraordinarily clumsy way. Would you agree with that, Mr Thompson?
Jon Thompson: They failed on providing basic levels of customer service on the telephone, which led to this situation.
Q42 Helen Goodman: You think the only problem was the telephone handling. Do you think that was the only problem?
Jon Thompson: No. It is a pretty massive problem. I have tried to explain that to other members of the Committee in the opening question. To have a fully functioning system, it is perfectly fine to ask questions about whether your claim is legitimate or not and whether you are giving us all the information. “Could you give us some further information?” is a perfectly legitimate part of the process. You are trying to balance off accuracy of benefits and speed of benefit payment against lower fraud and error. You are trying to balance those two things together, are you not? But if you ask a claimant to provide you with information and they cannot get through, you do not have a functioning system.
Q43 Helen Goodman: That is also true. In an earlier Parliament I was a Minister at DWP and we had—and I think this still exists—a fraud hotline. If people thought that somebody else was fiddling, they could phone up and say so. What concerns me is that Concentrix have not said where they have been getting the data from that has made them feel that 324,000 people may not have been telling the truth about their status. I would like to ask you—this is like a two‑parter—was that coming from the hotline for people reporting fraud and, if not, who are the third party data providers that Concentrix have been employing to do that part of the work?
Nick Lodge: The answer to the first part of the question is no; it would not have been people phoning a fraud hotline. The answer to the second part of the question is that Concentrix would have used, I am sure, various sets of data, including credit reference agency data from the well‑known suppliers of credit reference agency data, to understand the financial links between people, for example.
Q44 Helen Goodman: This morning you have been very reluctant to name these people and Concentrix were very reluctant to name them before the DWP Select Committee. Given that these people are being paid using taxpayers’ money, would it not be reasonable for the taxpayer to know who it is who has been supplying this information that has turned out to be so unhelpful to so many of our constituents? Should we not be able to have a sight through all the way down to who has got these contracts?
Nick Lodge: Concentrix used more than one credit reference agency. I am struggling to remember the names of the agencies that they used, but they were household name credit reference agencies, as far as I remember. We can name them, if you would like to, or we can write to you with the names.
Helen Goodman: Perhaps you could check it out with Concentrix and write to us about that as well. Thank you very much.
Q45 Chair: Let’s go back to several points that my colleagues touched on there. First of all, could we have clarity? What was the total value of the savings that if the Concentrix contract had gone as originally intended might have been available? Clearly the contracting out was done because it was hoped that there would be some public benefit. What was the value that was expected at the time the decision was taken?
Jon Thompson: You will forgive me that it was several years before I took up the post as Chief Executive.
Chair: I understand. You do not need to give me a precise figure. Give me a ballpark figure.
Jon Thompson: In which case Nick, I am sure, will give you a more accurate figure.
Nick Lodge: The most recent forecast was the 405 figure.
Chair: 405—
Nick Lodge: The £405 million that was mentioned earlier. That was in the most recent budget forecast.
Q46 Chair: Is any of that going to be recouped?
Nick Lodge: Yes. The last estimate that I saw, which will be updated to take account of the reviews that are going on and so on and so forth—so it probably will reduce—was about £270 million.
Q47 Chair: This is what I want to get on to, because the decisions that appear to have been announced in the House yesterday whereby the contract is going to be set aside, if that can be achieved, and this is all going to be brought back in-house is going to require some reorganisation for you to get on with the work. Work that had originally been taken away from these staff is now being given back to them.
Jon Thompson: Sorry, there is a slight misunderstanding that at some point this was an outsourcing and that it is coming back. This was never an outsourcing in the first place. It was additional capacity. It is not that it has gone out and is coming back.
Q48 Chair: Okay. It is even worse in a sense from the point of view of the burden that this will now place on you. There must be an Exchequer cost.
Jon Thompson: Subject to us reaching a final agreement, the staff working on this will transfer to HMRC and work for us. Yes, there will be a cost in that, because they will then become civil servants and we will pay their salary.
Q49 Chair: What is that cost?
Jon Thompson: We will pay their salary, but we would otherwise have been paying Concentrix.
Q50 Chair: What is that cost?
Jon Thompson: I do not have to hand what the difference is between paying Concentrix and paying staff. I do not have that to hand.
Chair: That would be of interest to the Committee too. These are core management numbers that matter a good deal in all this.
Helen Goodman: I have a question.
Q51 Chair: In a moment perhaps—I just want to clarify one other point. You said, I think, Mr Lodge, that 20% of cases have an error over the course of a year. Is that correct or did I misunderstand or mishear what you said?
Nick Lodge: No, you did not mishear what I said. We do an annual sample—a statistically valid sample—of quite a few thousand cases, and that statistically valid sample produces results that we publish each year that describe the extent of error and fraud in the tax credit system and break it down into various component parts. I was describing when I mentioned the 20% that that sample for the latest year—the tax year 2014-15—showed that broadly 20% of awards would have had some level of incorrectness in them.
Q52 Chair: Now let’s just go back to the answer you gave to Helen Goodman with respect to the evidence given by Philip Cassidy before the Work and Pensions Committee, where he said those are records that are flagged up as high risk from a HMRC standpoint. That is 1.5 million, which is a third. Are you saying those two numbers are apples and apples or are they apples and pears—the 20% and the 35% or 33% or whatever that number is?
Nick Lodge: They are slightly different things. The 20% is derived from a statistically valid sample that tells us the level of incorrectness in the tax credit system looking back to a particular tax credit year, in that case 2014-15. The 1.5 million would have been cases that exhibited some of the characteristics that generally can lead to an award being incorrect.
Chair: They are very similar.
Nick Lodge: We are absolutely not saying that they were all incorrect. We are simply saying that based on what we know, following several years of deep analysis, these kinds of cases can be wrong.
Q53 Chair: They are very similar. They are almost comparable on the basis of the description you have just given me, are they not? What I am trying to get at is what you seem to be saying is that a majority of the cases that are identified as high risk are indeed likely, on the basis of your past sampling evidence, to contain an error.
Nick Lodge: Certainly some of them would contain an error.
Q54 Chair: I know there will be some, but I am trying to get beyond some to a number.
Nick Lodge: We are trying to identify those cases that might be wrong and then do further analysis on those cases. In this case it was Concentrix doing further analysis on those cases, but we do much the same within HMRC as well. We do further analysis on those cases and select the ones that we think then on that further analysis are at risk of being wrong. Then, in order to be more certain, you have to talk to the claimant, contact the claimant and understand their household circumstances and any changes that might have happened. Then we go out and contact the claimant. There is a difference in that the 20% is what our sample tells us and the 1.5 million is: “Here is a big batch of cases, some of which we think are likely to be wrong.”
Q55 Chair: I am sorry to interrupt, but if we think that the sample is a broadly accurate sample, these are broadly apples and apples, not apples and pears, and therefore it is reasonable to say that a majority of the cases that are identified as high risk are likely to contain an error, which was the question I asked you a moment ago.
Nick Lodge: You could draw that conclusion.
Q56 Chair: I am asking whether it is reasonable to draw that conclusion, and if not, why not.
Nick Lodge: As I have described, we look at the characteristics that are likely to lead to incorrectness. It is only when we check the individual claim that we can know whether it is incorrect.
Q57 Chair: I am not asking for precision. We are dealing in percentages of very large numbers here. We have got that far. Now I want to ask in somewhat more detail the question that I opened up with in a very gentle way to Mr Thompson. Of those errors, what proportion do you think are fraud?
Nick Lodge: We categorise when we do the sampling exercise that I described, which is published. We do break down and categorise into categories of error and categories of fraud. It varies year to year, but broadly it is about half and half.
Chair: You are saying to us that 10% of the total cases are fraudulent.
Nick Lodge: The 50/50 that I am mentioning is by value rather than by number, because quite a lot of the 20% will be quite small errors. The total value of error and fraud in the system in 2014-15 was 4.8%. That is 4.8% of the £28 billion that we have paid out.
Q58 Chair: Is the average fraud bigger or smaller than the average error?
Nick Lodge: That, I am afraid, I would not—
Q59 Chair: This is hard pounding here. I am trying to get to some ballpark numbers so we can establish as a Committee how much is error. I am not talking about exact numbers on which you will be held to hell and high water. I am just asking you to give us some guidance on what proportion is error and what proportion is fraud by number.
Nick Lodge: It does vary, but broadly it has been around 50/50.
Chair: I am asking for averages over a run of years.
Nick Lodge: I would have to go back to the detailed figures to give you that.
Chair: Right. Okay. I will abandon this search for the time being, but we may have to come back to it.
Q60 Wes Streeting: Just very briefly I want to come back to mainly the contractual issues. I just want to pick up this point that Helen Goodman raised about cash flow. Jon Thompson, you were very emphatic in saying that this was not about cash-flow benefit and that was not a consideration for HMRC. At what point did HMRC consider the cash-flow disadvantage, and in some cases harm, that would be caused to people who have been caught up in these poor judgments and who rightly expect when money is owed to them that it will be repaid in full and up front in much the same way that HMRC expects money owed to HMRC to be paid?
Jon Thompson: We do consider that in the mandatory reconsideration process. Certainly in a number of cases I have signed off in response to MPs we offer some of the claimants a small amount of compensation up to £100.
Q61 Wes Streeting: You have said you have given an undertaking to the Work and Pensions Committee to look at it, but it is still the case, is it not, that if you are owed £1,500 you might get a lump sum to some extent up front but then more money trickles through. Why is it not just an automatic repayment of money owed?
Nick Lodge: I can explain the way that usually works. The majority of people will get a lump-sum payment. As you know, the tax credits system is an annual system. At the point somebody comes back into payment and we know exactly what their household circumstances are, the system will do a calculation and it will say what their annual entitlement is based on what we know about the household.
For illustrative purposes, assume we are halfway through the year. It will then look at the amount that has been paid out so far and it will compare that with the amount that should have been paid out so far on the basis of the current entitlement. If the figure that should have been paid out so far is greater than the figure that has been paid out so far, that amount is paid out, so people will get that lump-sum payment.
Of the cases I have had looked at, and I have looked at a few, where people have been, as you rightly describe, concerned that they have not received the amount of money that they should have done in a lump sum in a repayment for the weeks that they have missed, generally what has happened there is that their circumstances have changed. We are at a particular point in the year where, because tax credits are annual and based on annual income, if people’s income goes up, their entitlement goes down.
Very often what is happening is a mixture of things, but we do not withhold an amount that people are entitled to part way through the year. If they are entitled to that amount, it does get paid out, certainly on the cases that I have looked at.
As Jon has said though, we have heard the same messages that you have heard—that people have found that difficult and in some cases feel that they have not had the amount paid out to them in that lump sum that they should have done. We are looking at that to see if there are other things that we should be doing, but the system generally will pay out the amount that people had been entitled to at that point.
Q62 Wes Streeting: I will be honest. I am not sure I am clearer now than I was when I asked the question, and I understand the complexity of measuring annual income and changes of circumstances. We know that is the heart of the challenge, but it seemed to me that Helen Goodman painted a scenario that was accepted and HMRC has given an undertaking to the Work and Pensions Committee to look at it. Is there a problem that is being looked at or is there not a problem, in which case why are you looking at it?
Nick Lodge: We absolutely are looking at it, because we have heard from claimants that they have found it difficult to believe that they have had the right amount of money. We are looking at that because we want to make sure that claimants have the right kind of experience, and if they need a different system, for example to be put on weekly payments, then that is what we will look to do. What I am trying to get across, and I am sorry if I have not been clear, is that generally the system will pay out the right amount of money in accordance with somebody’s entitlement. If they have not had the right amount of money up to that point in the year, the system will pay a lump-sum payment.
Jon Thompson: Do you want to know what the specific problem is that Work and Pensions put to us?
Wes Streeting: Yes.
Jon Thompson: The specific problem that was put to us was if you are missing, to use your situation, the eight weeks and then we add the eight weeks on to the remainder of the year, you get a higher award and then that bumps into other aspects of the benefits system, notably housing benefit. If what that then means is you are getting higher tax credits because it is essentially on a gap period and now you get less housing benefit, all round you are worse off. That is the specific problem that we were given to look at.
Q63 Wes Streeting: That I understand. Right. I have got you. Let’s go back to the contractual issue and contract monitoring, because one of your duties as the contract awarder is to undertake proper contract monitoring and to be satisfied that all the delegated functions you have given to Concentrix are being lawfully and properly carried out. In that context when, if at all, did your monitoring notice a material change in Concentrix’s performance in that respect?
Jon Thompson: The problem with the telephony service occurred in the week commencing 15 August. I was alerted on Wednesday 31 August and then we had a conversation about it on Monday the 5th. It took just over two weeks to get from it happening on one day and then it eventually flowing to us. One of the lessons we need to learn from this is if something like this were to ever occur again how quickly you get from the operational level ultimately to the decision-makers of the organisation.
Q64 Wes Streeting: That is very clear and precise about the telephone issues. Were there other aspects of contract performance that were of concern and, if there were, what aspects of the contract performance did concern you and when and how did that arise as a concern to you?
Jon Thompson: Do you want to talk about broader aspects?
Nick Lodge: Yes, certainly. As I described earlier, we would monitor several hundred cases generally, and check several hundred cases, several hundred decisions, routinely every month. If there were issues with those, we would feed those back. We put a correction plan in place in the very late part of 2015 through a couple of months into 2016 to deal with and make sure that quality was improved. We had regular weekly performance meetings with Concentrix to ensure that those standards were adhered to. There were from time to time issues that were addressed in the way that I have just described, and there were from time to time issues, as John has said, with the call handling. August was, as Jon has described, the most recent. There had been some others where performance had varied and we had asked Concentrix to put that right.
Q65 Wes Streeting: You have given me a bit more clarity about the telephone aspect. Were there other aspects? At the moment it sounds a bit fuzzy on the other stuff. Were the concerns restricted to telephone-performance issues or were there other concerns, and if so what were they and how and when were they identified?
Nick Lodge: We identified some concerns with hitting the quality standards and put in place a correction plan with Concentrix.
Q66 Wes Streeting: Is that quality standards in terms of judgments or quality standards in terms of customer service?
Nick Lodge: That was in terms of the quality of the decision-making process, which is not just whether the decision is right or wrong; it is also whether the right things are recorded on the system. There is quite a wide list of things that somebody working these cases would have to do to absolutely get everything 100% right, so we were working with Concentrix to make sure those standards were upheld.
Q67 Wes Streeting: Are you satisfied that Concentrix have been applying the law correctly, in particular with the burden of proof being on them to demonstrate that a tax credits claim is wrong?
Nick Lodge: As Jon described earlier, there are two different aspects to that, but broadly the answer is yes, we were, and that is why we were doing the quality monitoring.
Q68 Wes Streeting: With the quality monitoring you were satisfied that Concentrix were applying the law correctly, particularly in terms of the burden of proof.
Nick Lodge: Yes.
Q69 Wes Streeting: Okay. Coming back to the points that were raised both by the Chairman and by Helen Goodman, particularly given the high proportion of mandatory reconsiderations that were found in favour of the claimant, do you think it is fair to say that Concentrix have taken a blanket approach in contacting tax claimants to ask for proof of eligibility rather than proceeding on a more evidence-led basis?
Nick Lodge: Concentrix, as described earlier, would have done their own analysis. This was certainly not, as has been mentioned in some quarters, a blanket fishing expedition. No, they would have written to people where they thought there was likely to something wrong with their award, and in some of these cases unfortunately, for one reason or another—perhaps claimants do not respond in time—the mandatory reconsideration process kicks in at that particular point.
Certainly Concentrix would have been writing to people where they had evidence or a reason to suggest that there might be something wrong with the award. That is why they would have got in touch with the claimant to find out the precise details.
Q70 Wes Streeting: I am slightly troubled by this. I will come back to it in just a second. Looking at the evidence, you told the Work and Pensions Committee that 73% of mandatory reconsiderations conducted by Concentrix were found in favour of the claimant, but in your letter to our Chairman of 21 October you said it was 95%. What explanation do you have for the difference in those two numbers?
Jon Thompson: Different periods.
Wes Streeting: They cover different periods.
Jon Thompson: You are missing a fact somewhere in this, which is the rate of mandatory reconsiderations versus the caseload. You have gone from “what are the cases?” straight to the mandatory reconsideration rate being 90% upheld in relation to the customer. There is a middle stage in there, which is the rate of people asking for a mandatory reconsideration. You can then draw some inference about what that rate is.
In relation to high-risk renewals, the rate of people asking for a mandatory reconsideration is around 8% for Concentrix and it is about 6.6% for us. They are not dissimilar. You could draw some inference about the 92%. If I make an enquiry and then you make a decision and 92% of people do not question that, you could draw the inference that those are probably right, but you have leapt straight to the 90% error rate. I do not think that is fair.
Q71 Wes Streeting: We will come back to that in just a second, because that is more looking to the future. I have a question that you can reflect on before we come to it, which is about whether people understand their entitlement to mandatory reconsideration and in what circumstances. We will come back to that.
I accept the premise of what you are saying, which is that overall, as a proportion of tax credit claimants and in terms of decisions, a relatively small percentage of people are asking for mandatory reconsiderations in the first place. I cannot understand how we can have—whether it is 73% in one period or 95% in another period—a huge percentage of mandatory reconsiderations upheld in favour of the claimant.
Jon Thompson: Yes.
Q72 Wes Streeting: My next question would be how many people who have had a judgment made against them who have not applied for mandatory consideration might also qualify, frankly, given those numbers? I do not understand how, with those sorts of percentages, you could possibly argue, as you just did, that this is not a blanket fishing expedition. When you have such a high proportion of mandatory reconsiderations being upheld in favour of the claimant, how could that not be a blanket fishing expedition?
Jon Thompson: You, again, for the second time, missed out the middle bit. If you write to 100 people and you make a decision to change somebody’s tax credits and nine of them write back and say you got it wrong and nine of them are right, what do you conclude about the 91 people did not?
In order to get the result that the Chairman was asking about in relation to delivering benefits to the Exchequer of £270 million, you have to make some assumption that some people did not respond to the enquiry on the basis that the enquiry was correct—that they were living with somebody and therefore the claim was not correct.
In relation to your question, you are trying to balance off two things. You are trying to get the claim accurate for those people who are entitled as quickly as you can, but you are also trying to remove from the system people who are clearly trying to defraud the taxpayer, and you are also trying to correct those people who make an error. Given the complexity of the system, which is suboptimal and incredibly complicated for people, it is relatively easy to make an error. This system should not have been designed in this way in my opinion.
Q73 Wes Streeting: Going back to the premise of your response, what I would not do is draw a conclusion about the 90‑odd per cent of cases you are talking about who do not apply for mandatory reconsideration. I would ask what percentage of the remaining 90% or so are people whose decisions have been accurate. If that is a large proportion or all of them, that is fantastic. My question would be, given the significant proportion of people who have their claims upheld, how certain can we be that the judgment for those who do not apply for mandatory reconsideration is right given these numbers? That would be my question and it may be worth reflecting on what HMRC does to provide further reassurance about them.
I want to move on to the allegations, such as those cited by Fiona Mactaggart in the Westminster Hall debate on 18 October, that it was HMRC who pressed Concentrix to pursue cases on flimsy evidence. In terms of the work that has been brought back into HMRC and will no longer be conducted by Concentrix, can we, as Parliamentarians, and customers be assured that we are going to get a better service?
It looks to me, based on what Fiona said in the Westminster Hall debate and the percentages we are looking at this morning, that HMRC is farming out the difficult stuff to Concentrix and putting in performance targets that incentivise people to pursue claims on flimsy evidence, and that is one of the reasons why we have such high mandatory reconsideration success stories within Concentrix and it is coming back in-house. I am not clear about whether HMRC will take a more evidence-led and intelligent approach to decisions or whether you will take the same cavalier approach to pursuing claims on flimsy evidence if you were the people providing the flimsy evidence.
Jon Thompson: I am slightly struggling with the binary nature of your question. We already run a tax credit system with enquiries for the majority of claimants. This contract added in that capacity. There already is an existing tax credits system within HMRC that is significantly bigger than that run by Concentrix.
Q74 Wes Streeting: I understand that. My question was more about whether you are chucking the difficult stuff out to Concentrix, and it is not going to be any better for those people coming back into HMRC.
Jon Thompson: I do not believe we are.
Nick Lodge: No, Concentrix did work on particular areas. We categorised tax credits, risk of error and fraud, into half a dozen different categories. HMRC does work across all six, including the same ones that Concentrix does.
Q75 Wes Streeting: But that is part of the problem, is it not?
Nick Lodge: We have been doing that work for some years.
Q76 Wes Streeting: Forgive me for interrupting, but that is part of and potentially at the heart of the problem with Concentrix. It goes back to what Helen Goodman was saying about the way in which your own system flags potential for fraud or error. I thought this was quite shocking when I read the transcript. Fiona Mactaggart said in Westminster Hall, “The director of Concentrix told me: ‘HMRC provide Concentrix with the claimant cases that they believe qualify for review.’ So the source of the evidence is HMRC. He continued: ‘These cases are selected by HMRC based on its own internal system which flags where there may be the potential for fraud or error. There were 1,497,000 cases provided from the Authority based on their initial assessment of risk or error and fraud.’” Concentrix is getting a hammering, quite rightly, for their poor telephone customer service, which is the problem that you have acknowledged this morning.
This is the third Select Committee hearing you have been to in the last few weeks, but it seems to me you think there is a telephone problem and that is what you are trying to solve. I do not think that you are acknowledging that there is a judgment problem here. There is an evidence problem here, and that is you, because you are providing Concentrix with these cases to investigate based on your system. Mandatory consideration success rates are huge. Your evidence is really flimsy and people are suffering as a result. That is not Concentrix’s fault; you are giving them the evidence.
Jon Thompson: Let’s try to deal with this in two parts. This is not a telephone problem. It is a telephone problem but it is not a telephone problem. It is a problem of making the correct award to benefit claimants and getting them back into payment. It is perfectly logical, to repeat my evidence, to ask a question, ask you to give us some further information, but if you cannot get through, you do not have a functioning system. That is basically manifest in the telephone system.
If you can fix that, you can have an interaction about what evidence you need, who it is that you think I might be a partner of and so on. You can then resolve your claim. You have to fix that problem, because it is right now and then it leads to the very human stories that were shared by Helen Goodman.
Separately from that is this question of our giving a certain amount of information to Concentrix. But the contract here was to go to the market and seek a third party who could bring you some capability to take the information we give them and apply additional information, analytics, other third party information, to be able to winnow that number down. That leads you to some of the questions that the Chairman was asking. You start with a population of X; further information is added by Concentrix, because that is the expertise that was put out to tender, and then that leads to the enquiry. It may well start with a population of cases from us. However, what was tendered out was Concentrix applying their analytical skills and third party data to reduce that and then make the enquiry. It is not a situation where we give Concentrix 100 cases and they chase 100 cases. It does not work like that in the slightest.
Q77 Wes Streeting: No, not at all. You are giving Concentrix 1,497,000 cases, based on your initial assessment of risk, error and fraud, and expecting them to do some further work around it. I do think Helen Goodman had a point when she pointed out that 1.5 million cases as a proportion of overall tax credit claimants is pretty huge. Some of your initial assumptions might need some work. I also think Fiona Mactaggart has a point when she says that we are beating up Concentrix about their judgments but the evidence trail starts with you. I am not clear and have not yet heard an assurance as to whether, when those cases that Concentrix would have been dealing with are with HMRC, the quality of judgment will be any better. It does not sound to me like you accept that there is a judgment error.
Jon Thompson: I thought I was conceding some of the ground. If we give Concentrix 1 million cases and they follow through on 200,000 of them, it is for them to make the decision about what other evidence they have between the 1 million and 200,000. That is exactly why the Government went to the market to bring somebody in. You could say, “Was the 1 million right in the first place?” That is definitely something we could have a look at, because we ourselves do generate other cases that we ourselves pursue. This is not a binary thing in which the tax credits system in this particular circumstance is entirely about Concentrix and what happened with Concentrix. There are other issues about the tax credit system that are suboptimal. There are other points about us having to accept responsibility for what has happened. I am not saying this is 100% a Concentrix issue.
Q78 Chair: “Suboptimal” does not really fully cover it, does it?
Jon Thompson: I could use some other, more colourful language if you would like.
Q79 Chair: Have a go.
Jon Thompson: It is deeply flawed, in my opinion, and of course the Government’s long-term solution is to stop tax credits altogether.
Wes Streeting: I will just ask two more questions. Thank you very much for your indulgence.
Chair: Two quick questions, Wes, if you would.
Q80 Wes Streeting: The first is: how do you respond to allegations that Concentrix errors resulted in multiple breaches of the Data Protection Act? Are you looking at that? The final question again comes back to this 1.5 million cases you are providing to Concentrix. I just wonder how it is that HMRC has so much money swishing around the system that you can waste public money giving a quarter of tax credit cases over to Concentrix for investigation on the basis of a high level of risk of fraud or error.
Q81 Chair: Perhaps you could just note those questions on the back of a piece of paper and we will come back to them right at the end of the hearing. They are very big questions. I know George has an engagement in Scotland he has to make.
Wes Streeting: I do not know how as a proportion you can contract out that sort of work.
Jon Thompson: I will give a written answer on the second, but on the first, yes.
Chair: We will come on to it in a moment.
Q82 George Kerevan: Tell us about these famous analytics that Concentrix supply that you did not have. What were they?
Nick Lodge: Concentrix were engaged to apply some analytics using whatever data they chose to use, in order to identify cases that they thought were at risk of being wrong. I would need to write to you on the detail of exactly what they did, but that was what they were engaged to do. We obviously do our own internal work, but they were engaged to add to that and to do their own analytical work, including using credit reference agency data, as I described earlier.
Q83 George Kerevan: So, Concentrix was employed to go out and find other information and they were sending letters to taxpayers seeking tax credits; I will not call them clients or customers. Is there not a possibility that, in seeking all sorts of other written evidence from taxpayers, those taxpayers might have initially been worried that they were having to give away private information to a private company and were reluctant to do so, or they did not do so because they thought it was fraud or a phishing exercise? Do we not think that was a fundamental problem with outsourcing and the way you went about it?
Jon Thompson: It is theoretically possible. The letter is from HM Revenue and Customs and Concentrix, but I can totally understand that some people might think, “This is from a private sector company and it might be phishing.” I can understand that some people might do that. I have got the letter in front of me. It does try to give some people some reassurance that, if they are not sure, there is a place they can go to to give themselves some reassurance that this is a legitimate letter on behalf of HMRC.
Q84 George Kerevan: I can absolutely assure you that I had quite a few people coming to my constituency office saying, “Who is this company Concentrix?” Some people are intelligent enough to google, and discovered it was an American multinational, and wondered what an American multi-national company was doing asking them for very private information. So, now that you admit that it was a possibility, in any future outsourcing will you take that into account?
Jon Thompson: I already gave the Committee an undertaking that we will not be going back to the market for this kind of work.
Q85 George Kerevan: Can you define “this kind of work”?
Jon Thompson: Let me be clearer then. We will not be going back to the market to seek a third party to help us in any way with the tax credits system.
Q86 George Kerevan: Given the tax credits system is on its way out, will you go to third parties in the market to assist you with universal credit?
Jon Thompson: We are not responsible for universal credit.
Q87 George Kerevan: Are there any aspects of HMRC business in which you would involve third parties?
Jon Thompson: There already is a vast array of work where we work with third parties, but not in what you might call front of house, to my knowledge.
Q88 George Kerevan: I do not know what you mean by “front of house”.
Jon Thompson: I mean in terms of directly interacting with customers. I think this is the only area. I can double-check.
Nick Lodge: I think the only other area where we would see taxpayers or claimants come into contact with third parties would be debt collection. There is a Government debt collection arrangement in place, of which HMRC is a part.
Chair: You began that reply with “I think”. I am confident that we would like to know a firm answer to that, because you do not seem sure. It is a very important question.
Nick Lodge: We can be sure and give you a written response.
Q89 George Kerevan: As far as you are concerned, in your tenure of office you are now ruling out replicating the kind of experience we have had with Concentrix. You will not go to the market for third party, private companies to interface with taxpayers.
Jon Thompson: In relation to tax credits, no. You know what the Government’s long-term strategy is—for tax credits to be replaced by universal credit. I believe that should be complete by 2022. There is frankly not much point in us doing that, particularly if the staff are going to come and work for us.
Q90 George Kerevan: That is very clear. I would like to get on the record the fact that third parties are being used in universal credit. My constituency is one of the first to have the full service roll-out of universal credit. We are having much the same problems with other companies. The problem, as far as the Government is concerned, has not gone away. It may be off your agenda.
Jon Thompson: You would need to ask Sir Robert to come in and talk to you about that.
Q91 George Kerevan: Indeed; I will excuse you on that one. You did have a contract review with Concentrix in October of last year. Was it not apparent at that point that there were problems, both with mandatory and telephone issues?
Jon Thompson: That is before my time, but perhaps Nick could answer.
Nick Lodge: I was in place at the time. The review led to some changes to the terms of the contract, which was published in the normal way, in October 2015, as you rightly say. The issues that we have seen recently and have been discussing were definitely not as apparent at that particular point. There had been some variation in service levels on the phone, but not with the mandatory reconsideration, as you mentioned. We were obviously satisfied at that point that the changes that we were putting in place would lead to improved performance, as indeed they did.
Q92 George Kerevan: Something must have happened between October of last year and now, because now that we are looking at the mandatory reconsiderations, there is a 90% failure rate. I accept what Mr Thompson said—that that is a small subset of the larger figure. But common sense would tell you that, even in a small subset, if a contractor being paid by the taxpayer is getting it 90% wrong, there is a fundamental issue there. However, you did not pick that up in October last year.
Nick Lodge: No, because I do not think the numbers were the same before October last year. HMRC are dealing with all of these mandatory reconsiderations now, as we mentioned earlier. Talking to the staff who are dealing with those mandatory reconsiderations, a lot of them arose because, for one reason or another, people had not responded with the evidence that had been requested by Concentrix. That may have been because people could not get through on the phone to explain their circumstances, and then used the mandatory reconsideration route. Mandatory reconsideration is an opportunity, and sometimes the first opportunity, to consider that evidence, but the numbers were different before October.
Q93 George Kerevan: I will just clarify that. You are saying one possibility is that before October Concentrix was getting it as wrong as often; people could get the problem resolved on the phone, but mistakes were still being made by Concentrix. However, because post-October Concentrix was not keeping up keeping up with the phone calls, the same, generic problem was then reappearing in the mandatory reconsiderations.
Nick Lodge: We did not see the same numbers around mandatory reconsiderations as we have just been describing—the 90%. I was suggesting that one potential reason for people seeking a mandatory reconsideration is that if they cannot get through on the phone to describe why indeed they are eligible for tax credits, they might then submit a mandatory reconsideration instead. Sometimes that is the first time that you then get to see the evidence. That is what I think some of our staff are describing to us. So, what we saw following October 2015 was a period of reasonable performance and improving performance, until, as Jon described earlier, we got to August.
Q94 George Kerevan: What happened in August?
Nick Lodge: In mid-August, the customer service that Concentrix were able to deliver dropped very significantly, and that is when HMRC took the decisions that it did. Those were not to extend the contract, as we have described, but also to take back the 181,000 cases that Jon described and take over the handling of the mandatory reconsiderations to improve the service for customers and to make sure claimants’ interests were put first.
Q95 George Kerevan: I understand. So, was this wilful on the part of Concentrix, in the hope that they could get out of the contract?
Nick Lodge: I do not think so, no.
Q96 George Kerevan: Were they just incompetent? Did they not put enough resources in? Why did this suddenly appear?
Jon Thompson: Our hypothesis is that it is about people—absolutely the hypotheses you have just put. It is about numbers of people put on the telephone lines. There is extensive research here about how queuing theory works. We can send out 100 letters; how many people probably are going to respond, and therefore how many people do you need to have on the telephone system? If you want me to go into that, I am very happy to do so.
In the end, though, the fundamental issue here seems to be that there were not enough people on the telephone lines to be able to answer the calls. We need to be transparent with you. HMRC went through this problem in 2014 and was subject to an extensive National Audit Office review, which went to the Public Accounts Committee in July. That was fresh in my mind when we came in to this situation. I know that the first thing you have to do is sort that out and get yourself back into a stable situation, because queuing theory says if somebody cannot get through they will carry on ringing and the queue just gets worse. You need to deal with it really quickly.
Q97 George Kerevan: I appreciate that. Concentrix is one of the world’s biggest outsourcing service providers—not a small company—and a major global player. It constantly presents itself in its advertising as being super efficient. I am just trying to grasp why, in what is clearly a flagship contract in the UK, it got it so wrong. You are ruling out the prospect that Concentrix began to realise it was under a lot of pressure—its margins were under pressure and it was under a lot of public pressure—and then decided to get out of this as fast as possible to save its reputation. Would you rule that out?
Jon Thompson: The only person who could answer that question, to be fair, would be somebody from Concentrix. I am sure that was not their intention.
Q98 George Kerevan: The NAO is conducting an investigation into the performance of Concentrix and how it was managed by you. What would you hope to get out of that in a positive way that would help HMRC in the future?
Jon Thompson: I have a working view about what the lessons to be learnt are. I will happily run through the five lessons I think we need to learn. We discussed this with the Public Accounts Committee yesterday, and what we are looking for from the National Audit Office is whether my working theory of the five lessons to be learnt is sound or not, and whether there are more lessons to be learnt. I will happily run through the five quickly, if you want.
George Kerevan: Yes, very briefly.
Jon Thompson: First of all, do we give enough priority to the sensitivity of customer service? Secondly, what do we need to learn about contingency planning?
Q99 Chair: Well, the first one is “not enough”. The answer to the second one is clearly you have not done enough.
Jon Thompson: I am grateful for your answers.
Q100 Chair: Are you disagreeing with them?
Jon Thompson: It is my view that we have not given enough priority to customers.
Chair: It would be helpful just to say so.
Jon Thompson: In relation to whether we have a thorough enough assessment of contingency planning, on the face of it we do not. Was the speed of escalation to decision-makers such as Nick and me quick enough? One would say, “Probably not.” It was fairly quick and then we acted, but that question is definitely there. The fourth point is the ability of third parties to understand the subtlety of delivering public services. The best of the private sector absolutely do get that, but is that the situation here?
Q101 Chair: Is “no” your tentative conclusion to that, or “yes”?
Jon Thompson: I will put that on the table, because I think it is a good question, but yes: my theory is that the subtleties of doing this possibly were not understood well enough. However, that is a theory. The fifth point is the question about incentives. Is an incentive-based contract, in this kind of area, the right thing or not? It seems to me those are five points.
Q102 George Kerevan: What is your answer to the final one on incentive-based contracts?
Jon Thompson: The fact I am bringing it into the room obviously means I have got some doubt. Those are the five I put to Work and Pensions and that I am putting to you and that I put to PAC. The NAO are going to do that, and we will see whether they agree with me or not and undoubtedly give me some other views.
Q103 George Kerevan: All things considered, if you were starting afresh would you have entered into this contract with Concentrix?
Jon Thompson: It is best to avoid theoretical questions about the past when you were not responsible.
George Kerevan: It is a test of your judgment.
Q104 Chair: Putting it the other way round, those are hypothetical questions about the future, rather than the ones that we know the facts for, that it would be helpful to avoid. We would be grateful for an answer to that question.
Jon Thompson: I do not think if we were letting this contract in the future, and I have already told you that we are not, I would go with an incentives-based contract.
Q105 Chair: I have been passed a note by staff reminding me that getting VAT back by those who are taking things out of the country for export is conducted by private companies at the airports. Is that not a HMRC front-end activity that is contracted out? Is that not going to be quite large?
Jon Thompson: I would have to check. I was not anticipating you would ask me that question.
Q106 Chair: I am concerned, I have to say, about the vagueness of some of the replies we have received and the lack of clarity about some of the crucial numbers of size of fraud and error. The NAO are conducting an inquiry. It sounds to me as if you might benefit from doing your own internal inquiry in more depth so that you can arrive at much more detailed answers to the points that you yourself are making in response to George’s last question. Do you think that might be helpful?
Jon Thompson: It might be. We need to be really clear about what my priorities are. My priorities are to resolve the situation right now for as many claimants as possible. That has to be the top of our list. When we get to that point—we are nearly there but not quite—Nick and I need to sit down and work out what happens next. We already know there is going to be a NAO piece of work and they have already started. I am not sure whether there would be some further value in another internal piece of work. My priority is to sort this out for customers.
Q107 Chair: You say that is a job that you think is nearly done, but not yet. When is it going to be done?
Jon Thompson: We gave you the data up front. We would anticipate all of the incomplete cases should be resolved next week, if the outstanding ones with claimants come back to us. Mandatory reconsiderations will take a bit longer than that because we are still receiving them, but we are well within the performance metric of how quickly we can sort those out. We have still got 6,000 mandatory reconsiderations to go, but at the moment we are turning them around in roughly seven days, when the target is 21 days.
Q108 Chair: In a few weeks you should be able to sit down and start thinking about those and other questions. The Treasury Committee would find it helpful if you did sit down and think about them and give us the preliminary product of your own internal inquiry and thoughts by the end of the year. Would that be reasonable?
Jon Thompson: The National Audit Office has turned up and is undoubtedly going to do a thorough piece of work. I am not clear what the benefit would be of us doing a parallel piece of work when you presumably have access to the National Audit Office piece of work and, indeed, the National Audit Office could presumably come and give you evidence. I would happily do that with Sir Amyas if you were so minded to do that. I am not clear what additional benefit there would be in doing another piece of internal work together with the NAO.
Q109 Chair: I am surprised by that reply. I would have thought the very first thing you would want to do would be an internal inquiry so that you were clear on the explanations for the mishaps when you are asked questions about them by the NAO.
Jon Thompson: Let me have a think about it. I am not committing myself to producing an internal inquiry document in the same timeframe as the National Audit Office.
Chair: We are not at the point where we are requiring it but we may be. We would like you to think about it very carefully and perhaps come back to us on that point.
Q110 Helen Goodman: I will just come back to repayments when people have their tax credits restored and they get a repayment for past underpayment. Of course, when the claimant is asked to repay, they get a letter saying, “You owe HMRC £1,500,” or whatever the amount is, so it ought to be possible to do an exact mirror image when it is your turn to make the repayment and work out precisely what the number is. I think the DWP Committee, as well as raising the interaction with housing benefit, also pointed out to you that people had gone into debt and they wanted to repay their debts. That was another reason for sorting this out. Are people getting any payments for the distress that has been caused to them in addition to getting the entitlement that they should have had previously?
Q111 Chair: We should make this the last question. It would be helpful to have a clear reply to this.
Jon Thompson: I think Mr Streeting asked me that question. The answer to that is “up to £100”.
Q112 Helen Goodman: Is that an additional amount?
Jon Thompson: That is in addition to the amount due to them for the period that was missing.
Q113 Chair: It would be helpful if we could have an estimate of the total value of those payments and the number of people who have been in receipt of part or all of that £100.
Nick Lodge: We look at that on a case-by-case basis. When people come to us and complain about what has happened to them, we look at the redress that they might be due.
Chair: We are very grateful to you for your evidence. It is a sorry episode, but these things do happen. We understand that nothing is run perfectly. We are of course here, as is our job, to make sure that there is a high degree of transparency about what went wrong and to make absolutely sure that everything reasonable is being done to put it right. If it has been a somewhat uncomfortable morning—it is now the afternoon by a couple of minutes—that is why. Thank you very much for coming to give evidence to us today.