Public Accounts - Minutes of EvidenceHC 936

Back to Report

Oral Evidence

Taken before the Committee of Public Accounts

on Monday 17 December 2012

Members present:

Margaret Hodge (Chair)

Mr Richard Bacon

Stephen Barclay

Guto Bebb

Jackie Doyle-Price

Chris Heaton-Harris

Meg Hillier

Mr Stewart Jackson

Austin Mitchell

Nick Smith

Ian Swales

Justin Tomlinson


Amyas Morse, Comptroller and Auditor General, National Audit Office, Gabrielle Cohen, Assistant Auditor General, NAO, Max Tse, Director, NAO, David Clarke, Director, NAO, and Marius Gallaher, Alternate Treasury Officer of Accounts, were in attendance.


A commentary for the Committee of Public Accounts on the Work Programme outcome statistics
(HC 832)

Examination of Witness

Witness: Robert Devereux, Permanent Secretary, Department for Work and Pensions, gave evidence.

Q1 Chair: Given that we are short of time, on the Work programme, let me start by saying that last time we discussed this-I think it was in February-the first 10 minutes of our session were spent with your saying that the NAO had got its figures completely wrong because it said that your outcome predictions were over-optimistic. Do you wish now to apologise to the NAO?

Robert Devereux: No. We haven’t finished the contract, have we?

Q2 Chair: Those early figures, which are what the NAO said at the time were outcomes that you expected to collect or looked to achieve, I can give you quotes about them. You said you were very confident you would get to them. You have failed to achieve them, and I think you spent 10 minutes at the session-I have got the transcript here-attacking the NAO for not believing the figures that you gave us.

Robert Devereux: Two things. One, I am not going to say that the numbers here aren’t lower than we expected, so I agree with that.

Q3 Chair: Right. So you will apologise to the NAO, because that is actually what it said at the time.

Robert Devereux: No. With respect, what the NAO said at the time was that my calculations of dead-weight and performance were wildly optimistic and were never going to get there. And I went to some length to explain why it was that having gone through all the history-10 years’ worth of data-we believed that that dead-weight number would still rise.

Q4 Chair: The figures were wrong. The figures that came out are a third of what-

Robert Devereux: With respect-

Q5 Chair: That seems to me wildly optimistic. I don’t know what the definition of "wildly optimistic" is, but-

Amyas Morse: Sorry. May I be allowed to disclaim the use of the word "wildly"?

Chair: Okay. But I think that it is pretty wildly optimistic to achieve only a third of what you set out to achieve.

Robert Devereux: Okay. Let me try again.

These are contracts that allow every provider to spend up to two years with every single person that we refer to them. And the data that you have in front of you is drawn from July 2012, which is therefore one year-13 months-for the-

Q6 Chair: But you set the profile, Mr Devereux. We know we are early in the programme, but you set the profile. You said that by July 2012, you would have achieved over 100,000 outcomes and at our Committee meeting you said that the only figure on which we could all rely was outcomes. You said that outcomes were the only true measure of success; they are the only measure, apart from attachments, on which public money is paid. So it was the only true outcome-I don’t want to hear about other outcomes.

Robert Devereux: Correct. And at our previous hearing, we had no conversation-to my memory-as to the profile at which this would be achieved. The entire conversation was whether or not, at the end of two years with any one cohort, we would or would not have exceeded the 31% minimum job outcome, which every cohort is required to meet under my contract. As of today, what the Report says is that the first cohort have currently got to 10% of the people we have referred to them in June 2011; 10% of those people have secured a job outcome.

Q7 Chair: Have what?

Robert Devereux: Have secured a job outcome. Okay? 10%-9.8%, it says in the Report. That cohort has got to get-

Q8 Chair: What are you talking about? How are you defining a "job outcome" in this instance?

Robert Devereux: I am saying six months in work, which is what the payment mechanism is for most of the payment groups. The statistics that you are correctly quoting as being low are all to do with job outcomes.

Chair: Yes.

Robert Devereux: I am trying to explain why I haven’t yet given up on the conversation I had with the NAO the last time I came here.

Over the space of two years, what we said in this Report-I brought it with me, because I remember talking about it at some length-was that the minimum performance that we expected across all the contract was 31% of the people who were referred, from every single cohort, to have a job outcome like this.

What the Report says is that so far, for the June 2011 cohort, which is the very first cohort that went through this, they have now secured 9.8% job outcomes, and that is by July 2012. They have got until May 2013 to get 9.8% up to 31%. As the NAO observes, 42% of all those job outcome’s have been achieved in the last two months-

Q9 Chair: I think you are bundling us with figures again.

David Clarke: Sorry. 3.6% is the job outcomes.

Q10 Chair: Yes. 3.6%-that is what I’ve got. I couldn’t work out where you’d got even your 10%. The only figures that are properly validated are the 3.6%, which is the figure from the-

Robert Devereux: So maybe you would like to turn to paragraph 24-

Q11 Chair: You were very clear last time, Mr Devereux, that we couldn’t take any figures-because we all had them from our own constituencies-except for those from the Office for National Statistics.

Robert Devereux: If you would like to turn to paragraph 24, you will see that the number that I am quoting is reported by the NAO. It says, "9.8% of claimants referred to the programme in June 2011 had received a job outcome by July 2012."

David Clarke:Sorry, I think what we mean is that they have been placed into work. Is that what-?

Q12 Chair: Yes. Quite. That is why I asked the question. You were very clear that this was a Payment by Results system, which we all think is a good idea. You defined it as being in work for three months if you are ESA and six months if you are JSA.

David Clarke: Of the first cohort, you are talking about now. Yes.

Robert Devereux: I am. Thank you.

David Clarke: It is correct that the people who first started the programme-say the ones who started in June-the outcome rate for those people is now 9.8%, but just those people.

Robert Devereux: Just for those people.

Q13 Ian Swales: For clarification, Mr Clarke, in figure 4 where you have referrals, which is what the 3.6% is based on, is that all referrals? That might include people who were referred two or three months ago.

David Clarke: Yes, that is correct.

Q14 Ian Swales: So there is no way that they could have had a job outcome as defined by the scheme in that period.

David Clarke: That is correct. That is why there is this sliding scale as you go through the programme as to what you think the expected performance would be. Our estimate, using the Department’s criteria, would be that by the end of July you would probably get about 12% of people into work with job outcomes, but actually it is 3.6%, so it is about a third.

Q15 Ian Swales: In order to hit the 31% target?

David Clarke: That is over the whole period.

Robert Devereux: Exactly. If we go back to numbers, which we have now established that I am speaking correctly, the June cohort-

Q16 Chair: June, which?

Robert Devereux: June 2011, the first cohort through the run. I am trying to keep it simple because some of it is quite complicated. For the June 2011 cohort, every single one of those people who were referred, if you then asked how many had a job outcome by July the answer is 9.8%. Right? This programme will be a failure if they have not reached by 31% by May 2013. Right? You cannot tell now that it will be a failure. You can tell that is underneath the profile that we originally had. I am not going to say that it would not be better if it was higher. At the end of the day, the question is, will they actually get from 9.8%, based on outcomes to July 2012, to 31%?

Q17 Chair: What is going to change, given that you are so under profile? You are at a third of the level that you said that you would be at. Given that, what are you changing so that you remain confident that you can achieve your profile now into the future, July 2013?

Robert Devereux: Good question. Remember, the job outcome is the accumulation of six months in work; not necessarily contiguous but over the period of six months. With every passing month anybody who has been in work and gets another week’s work-they may be working over Christmas for all I know right now-they could actually be already at five and a half months and need two more weeks to get to six months. There is something about the cumulative nature of the process.

Q18 Chair: Are you changing nothing then? What I asked was: given that you are below profile at the moment and given there have been a lot of concerns about particular providers-we will come to that after the vote-and you have concerns, what are you changing to improve your performance so you get back into profile, so that you can maintain your confidence that by July ’13 you will achieve the performance you committed to?

Robert Devereux: The answer is that we are working with each of these providers to ensure that they all know what good performance looks like and what bad performance looks like. In that 9.8% you can see in the publicly available data-

Q19 Chair: You are not changing anything; you are just working with the existing providers. I will tell you the sort of thing that worries me. By July ’13-I may have my dates wrong-the providers will no longer get the attachment fee. Is that right?

Robert Devereux: Correct.

Q20 Chair: By July ’13 they are also going to have to start paying the discount rate to you. When do they start paying the discount that they got in the contracts?

David Clarke: She means the price, when they bid, they bid for a discount on the outcome fee-some providers.

Q21 Chair: When do they start paying that?

David Clarke: Some providers when they bid, they obviously bid with a discount on the outcome fee.

Q22 Chair: So that all starts by July 2013. If you look at the performance now, they are so under profile, and they are going to hit those two new financial constraints-the attachment fee and giving you back the discount-something has to change, in my humble view, to ensure that you achieve the performance you intended for this programme. I am asking a simple question. You say you are working with the providers. Is nothing else changing?

Robert Devereux: The most important thing that needs to change is their performance management of these contracts. There is no evidence in the data that this is somehow connected to local employment conditions.

Q23 Chair: But all of them perform under profile. There is not a single one that has performed to profile. Not one.

Robert Devereux: That is correct but, as the Report makes quite clear, it is difficult to work out what this two-year, ramp-up profile should be when NAO properly and diligently-

Q24 Chair: You can work it out statistically. I put it to you that, if in the real world something does not change, you will be back in front of us in July ’13 or ’14-if you are still in the job-

Robert Devereux: I intend to be.

Q25 Chair: You intend to be, good-you will be back in front of us then with some other excuse. I am astounded that you think your performance to date is sufficient to warrant the confidence, given the financial changes the providers face, to say that you will achieve things. I really am astounded.

Robert Devereux: I know you are, but I have not finished the answer. It is really important to make this point though-

Chair: Okay, but we will be going to vote, so do it quick.

Robert Devereux: The fact that this is a huge performance means that if, for example, people are off benefit, which they manifestly are, and many of them have had spells in work, then actually the lapse of time will dry this up. We may well have got-

Q26 Chair: That is nothing to do with the outcomes. It is nothing to do with the performance-

Robert Devereux: I am afraid it is, Chair. If more of them accumulate-

Amyas Morse: We think it is to do with the outcome as well-I have got to say. We are not encouraged by the fact that it is significantly off profile but if, for example, for some reason we do not know, the curve moves differently, it is not inconceivable that it could still-

Robert Devereux: I say that because some of the Report makes it quite clear that there are a number of reasons at the moment, which are depressing results; these things are not enough to make it go away completely but they are an important consideration. When you say that obviously I must change something, I am just asking you to bear with me because some of this may well turn out to be better performance management, and the later data may give us a better position.

However, that said, we will clearly look to see what else we can do, so it says in the Report that we are looking with our colleagues in BIS and the providers at if there is anything else we can do about the way we dispose of money on the schools provision. We are thinking, "Is there anything else that we can do to support the outcomes that you and I both want?", so it is not a question of simply saying, "Well, I am not going to do anything about it." The most important thing to do is to go back to these companies that have got responsibility for doing this and-

Chair: We will come back in 10 minutes.

Sitting suspended.

On resuming-

Q27 Chair: Where had we got to?

Robert Devereux: Whether I was going to change anything.

Q28 Chair: And you said you weren’t. Okay. Tell me one thing, because I am more sceptical than some of my colleagues round the table. I think it may well be that one or two of the providers cannot manage under your regime, which is why they are not performing. Do you expect any of the providers to go bust?

Robert Devereux: I think the previous Minister of State was quite clear that he thought there was a possibility that some may. The whole point of risk transfer is actually to transfer risk, so people who do not perform, as I explained at great length on a previous occasion, will not be able to make money out of this contract, and that will put them at a financial disadvantage.

Q29 Chair: But you are not, in any way, thinking that the discounts or any of that will go?

Robert Devereux: Why would I?

Q30 Chair: Because I think they will not be able to manage financially within the envelope you have given them to get the sort of performance you want.

Robert Devereux: This is a substantial break with the past in terms of the-

Q31 Chair: I do not think it is; it is a development from past policy.

Robert Devereux: It is a development that has the effect that no longer am I paying for anything less than an outcome of six months once I have got beyond year three. That being the case, I think it is possible that some people may find it a struggle to make ends meet, which is the question you asked me.

Q32 Chair: But you are not going to change the regime.

Robert Devereux: I am one year into this programme for all practical purposes, and it is a five-year programme. The last thing that you would ever want to do with contracts is start mucking around with them 12 months in.

Q33 Chair: Okay. So you are not going to change anything.

Robert Devereux: It could not possibly be value for money at this stage to say, "I tell you what, I’ll let you off some of the payments and some of the discounts."

Q34 Chair: But it is not value for money if you do not get the people into work. Can I just ask one final thing, and then I will hand over to someone else? On the same day you put out the statistics in a very odd way-we will come back to that-you also put out a DWP Work programme evaluation on your website. One thing it found particularly worrying relates to the parking of people. This comes out in the stats in the memorandum prepared by the NAO, and figure 3 on page 9 probably confirms what your own evaluation said. Let me quote you three sentences. First: "available evidence to date suggests that providers are engaging in creaming and parking, despite the differential payment regime". Secondly: "Those assessed as hardest-to-help are in many cases left with infrequent routine contact with advisers, and often with little or no likelihood of referral to specialist (and possibly costly) support". Thirdly: "some providers at least, took the view (perhaps surprisingly, given the design and remit of the Work Programme) that it was inappropriate for the hardest-to-help to be referred to their services at all." Those are three sentences taken out of your own evaluation. If you look at the figures on page 9, for the ESA ex-IB people down at the bottom, your performance is 0.2% achieved by July 2012. If you look at individuals, only 20 people have presumably been in work for three months. What are you doing about that? Again, you said you did not want hard-to-help people to be parked, but it is clearly happening, and your own evaluation shows it. What are you going to change to stop the creaming and parking?

Robert Devereux: At the risk of repeating myself, we have established a contract with a completely different method of payment from previously. Contractors will not be able to remain profitable unless they tackle substantial numbers of people well beyond the easy-to-help. At the moment, as you can see in the chart produced by the NAO, performance varies very significantly between providers. I am not about to change the nature of this in order to-

Q35 Chair: What are you doing to tackle the fact that performers are creaming and parking hard-to-help clients? What are you doing about that? It is your responsibility.

Robert Devereux: No. My responsibility, with respect, is to try to-

Q36 Chair: Okay. You have no responsibility to ex-IB, now ESA, claimants who want to work.

Robert Devereux: Our responsibility is to put in place the best value for money method for assisting-

Q37 Chair: You feel you have no responsibility to tackle a group who are clearly not being helped at present.

Robert Devereux: If I could finish the answer to any one question, we might know the answer. We have put in place a regime in all these different payment groups that tries to identify the right price that the Government should be paying for job outcomes, and we have made it perfectly plain that unless people are taking action right across that space, they are unlikely to remain profitable and remain in business. That is the clearest possible signal that I can give to the private sector to make choices.

That is the way the system was set up. It was not set up on the basis that somebody wise-a Minister, a civil servant-knew exactly how to do this, because all the previous systems, if you remember, were mostly unsuccessful in getting people back into work. We could be busy, but we will not be effective.

Q38 Chair: You gave evidence to us that the purpose of the contract and of the Work programme is to support people back into work in the most value-for-money way. That is the purpose, and it is a purpose shared by everybody in this room. If your own early evaluation and the statistics demonstrate that a particularly important group of people-those who are on incapacity benefit, whom we all want to get back into work-are not being properly supported despite your differential cost regime, that is what your evaluation and your statistics show. Are you telling me that you do not really care, as long as the contract pays, about what happens at the other end?

Robert Devereux: I am not saying I do not care, and the-

Q39 Chair: Well, what are you going to do about it, then? The evaluation shows you have got a problem.

Robert Devereux: I am going to come back to the fact that we are one year into this contract, and people are trying to work out the best way to do this. We are going through it with every one of these providers to understand why their performance is where it is in every one of these payment groups, to try to make sure that they understand what they can do that will improve that performance. I am a little bit coy about telling them that I know the answer, because as I said in the-

Chair: No, I am not suggesting-

Q40 Mr Bacon: Can you tell us what you have discovered so far? For example, it might not be where you want it to be, but if you look at ESG and Maximus, at the top of the chart on page 10, figure 4, ESG is achieving 5% and Maximus is achieving 4.6%. That is more than double the people at the bottom. What have you identified so far about what they do that is plainly working better than the people at the bottom? That cannot just be a statistical or geographical quirk. Presumably there is some difference in method that is having different results.

Robert Devereux: It is not, and it is the product of two different things, which is why my answer to the Chair is going to remain the answer. People are approaching this in different ways. They have a different process themselves, and they have different levels and competence of management going on.

In the first instance, I have deliberately not specified what the nature of the process should be. Some people have different processes. Secondly, different people in different areas are clearly more on top of their own internal MI-managing it, and making sure exactly who they are dealing with and what else they should do to maximise their outcomes-than others. I suspect that what you are seeing is a combination. Some have a better definition within the black box offered them, some have come up with a better process and some are managing it better. That is the thing that we will then be taking through.

We talked to all the contractors that we are using at the moment, because they are all required to produce performance improvement plans, to make sure we understand what is working well. I have not come with chapter and verse on what those are doing, because the data came out quite recently. That is why I am rapidly back in front of you, three weeks after the statistics were published, but that is the nature of what we are trying to do here. We are trying to understand whose processes are working and why those processes are better than others. Secondly, some people have different results in different areas with the same process, which will be down to the local management. That is exactly the same as most other things in life-

Q41 Chair: I have to say that I do not think you have answered the question on the hardest to help. Tell me one provider that is doing well with the hardest-to-help group, the ESA. Which provider is doing well there? I do not think there is one.

Robert Devereux: I do not think any of them are anywhere near where we would like them to be.

Q42 Chair: Quite. You have a problem there.

Robert Devereux: Yes, and the only way into that is to start to say, exactly as Mr Bacon said, "What are some of the people at the top of the tree doing? Is it possible that it is actually all to do with JSA? Could they translate it somewhere else?" People have to put more effort into understanding why they are getting good results where they are getting good results and learning from that elsewhere.

Chair: But nobody is getting good results on ESA.

Q43 Nick Smith: Mr Devereux, you talked about getting up to 10% or so of clients working for six months. We have talked about your end-target of 30%, and you are substantially off that end-target. You need a step change to get there. I hear what you say about it being early days, because this is very new. I also hear what you say about sometimes having to run programmes because they gain momentum over time, and they sometimes get better very quickly and you do very well. And I hear what you say about different places having different methods. I understand that complexity, but you will have a list of people who are clearly underperforming and are very unlikely to reach your end-target of 30%. At what stage are you going to have a review that cuts off the poor performers and says, "You really aren’t going to get up to speed, and you really aren’t going to deliver. We need to swap horses here." When are you going to do that, and how are you going to do it?

Robert Devereux: The contract provides, as the NAO clocks, for a concept of minimum performance. As I have explained to the Chair, the providers are allowed to work with every single claimant on the programme for two years. The point at which the first cohort reaches two years will be June ’13. We will have a very good indication when we get the data for the year ending March ’13. If at that stage people are underneath the minimum performance, the contract provides for me to be able to take action against them, including ultimately, if that is where the performance stays, taking the contract off them. If they are above the minimum performance but one provider is substantially better than the other, I have the capacity to switch work from the poorer provider to the better provider.

The short answer to your question is that, during the course of 2013, when they have had a chance to do two full years, which is what the contract provides, we will see whether they have reached 31%, which is the critical benchmark for happiness. If we get to that point, I will have some information that, within the contract, I can do something with. I have both the minimum-performance and the share-of-work clauses in the contract with which to operate.

Q44 Nick Smith: For some of them you will have a pretty good idea now that they are underperforming. What do you say to that?

Robert Devereux: What the Report records perfectly properly is that, for every single contract in place-it was 16 at the time of the field work, and it is all 40 contracts now-we have asked them to improve their performance, because obviously it is less than we wanted. It also goes on to say that, for seven of those providers, we have written letters to say, "This performance is really not acceptable, and you need to tell us urgently what it is you are going to do to improve it."

Q45 Chair: Are those the seven at the bottom of the list?

Robert Devereux: They are the seven who we have most reason to believe need to get our letter.

Q46 Chair: But are they bottom of the list?

Robert Devereux: They will be down that end of the market, but the point is that you are looking at data that are published; there is slightly later management information that some of those individuals have improved their performance.

Q47 Nick Smith: You have sent them a letter. What else are you doing?

Robert Devereux: The point of the work we are doing with all 40, and with the seven, is to get underneath, in a very forensic way, all the things I have just been discussing with Mr Bacon. I want to know these things exactly: "In respect of the brochures you’re doing, how do you know that this is working? Where is the MI that actually tells you this? If it is not achieving what you thought it was going to achieve, what are you doing about it? How can we share best practice?" In other words, it is very hands-on contract management of them against the process that they themselves have devised, and, in the event that it is self-evident that their process is not working, we say to them, "Look, this other process over here is working quite well."

Q48 Mr Bacon: You say it is very hands-on. Did you say you are just writing a letter to them, or are you meeting with them to talk about this?

Robert Devereux: Do you recall these people called provider assurance teams? I think that is the right phrase. I have teams of people working in the Department who are actually going out and going through in some detail what is going on in these contracts. So it is not just a case of sending a letter and hoping for the best; we are actually meeting them, looking at their data-

Q49 Mr Bacon: I have just realised that the acronym for provider assurance teams is "PATSY".

Robert Devereux: Well, there you go. I will try to call it something else next time round.

Q50 Nick Smith: Can you let us know a bit more about the challenge that you have given them beyond sending them letters, so that we can have some reasonable confidence that you are crawling all over these contracts?

Robert Devereux: Yes, sure. The first letter has literally come out. You have perfectly reasonably-since this has a big Government profile-had me back almost immediately. These teams are out doing this work now and have been for a while. I am very happy to write you a note. Why don’t I write you a note at the end of January, when they have actually done the work that they are currently engaged in, consequent upon these published statistics?

Chair: We will probably want our Report out before then.

Q51 Ian Swales: I guess one of the most damaging things about the media reports about all this is the sense that it is, if anything, worse than the do nothing option. Figure 2 tends to support that view. How do you account for the fact that the Department gave various trajectories of what might happen in one case if we did not do anything, yet we are performing below that?

Robert Devereux: Two things, I guess. One is, I am not going to renege on the fact that these are disappointing figures. We are below where the numbers should be. I have to say that projecting the profile to get from nought to 31% is not a science. I had a long conversation earlier about housing benefit and models. This is a lot better baked than it would have been in the housing benefit case, but we are still trying to work this out. But as I said, and as the NAO records, there are some structural things going on here. Some people got out of the blocks slower than we expected. They are on average taking a further month to claim-a month beyond what we thought-largely, by the way, because of the care with which we are requiring them to report so as not to get into the fraud and error that we have had with previous things. So, at the moment-

Q52 Chair: If you add those in, that is another 7,000.

Robert Devereux: I have not at any stage said that any of these things miraculously put the performance up.

Q53 Ian Swales: When you projected the do nothing option, presumably that was based not on a finger in the air, but on typically how the long-term unemployed enter the job market and so on.

Robert Devereux: Indeed. The only optimistic way to understand this would be if people positively thought they were giving people more skills to have a longer sustained period in employment over the period of the two years. Personally, I am not going to trouble you with that answer, because I am not sure that it is correct. At the moment, the figures are the figures. The point is that with no previous contract has this Committee been able even to know what the deadline is. We have at least built a contract, which started with 10 years of good econometric data. This is ahead of us now. With respect, having written a contract for five years and had private companies put real money up to cover their cash flow for the first couple of years, they do need two years to have a go at this.

Q54 Ian Swales: Going on to the companies, of course the most interesting data is not in here, and I do not suppose we will see it. It is on how they are doing in each area. These are the companies in total-

Robert Devereux: It is published, so you can go on to the website and dial up any combination of provider, contract, claimant type-

Ian Swales: I’ll do that.

Q55 Jackie Doyle-Price: It is not easy to navigate.

Robert Devereux: Okay. I did not say it was easy, because I tried myself and I found some data, but not everything. None the less, every possible combination of constituent, provider, data and payment group is out there. We are being totally transparent.

Q56 Ian Swales: That is some homework for me; thank you. Can I just ask this, in addition to what Nick was talking about a minute ago? To what extent are you looking not just at the outcomes, but at whether they are actually following what they said in their contracts? I ask because I live in a unitary authority area and one of the north-east providers has failed to set up at all in my area, yet still half the people are being referred to them. So they are having to travel out long distances to go and meet this provider. We have raised it with them, but they still have not done it. That provider, in their small print, said that 50% of their work would be done through the third sector, and I am not sure any of it is being done through the third sector, which is quite important given what this contract did to the third sector in lots of regards-not just council operations but the third sector charitable organisations, and so on. Are you doing any work to assess whether people are actually doing what they said they would do?

Robert Devereux: In the bidding process, we did ask people to set down what they were doing. We have essentially determined the bids on the basis of people who have got a plan, and then looked at the financials, and that is where you get it from, this discounting.

The provider assurance teams should be looking at the extent to which people are actually doing what they said they would do. But can I take that example away, because I do not know the answer to the question about-if they said they were setting up, what in practice is the difference. It is complicated-

Q57 Ian Swales: In their bid they named organisations that they had never even spoken to and have not spoken to since, either. They named organisations that they were going to work with. There is some poor practice there, in terms of how they contracted.

Especially given these organisations are likely to be short of money now, are you aware of their either gaming the system or not helping people in the way that was within the spirit of what they originally contracted, to save money?

Robert Devereux: Not to save money, no. I am not aware of that. The evidence-

Q58 Chair: Creaming and parking are not to save money?

Robert Devereux: No, I took the question to mean newly seeking to save money based on these results. The parking is what the parking is at the minute.

I am afraid that I am going to come back to the rather simple fact that they took these contracts on in order to deliver on outcomes I wanted, right? If they cannot do it at a price which we have established is a fair price for the taxpayer, then we do indeed have a problem. To come back to a point that the Chair made in another forum, I do understand that, at the end of the day, we do want to get long-term unemployed into work. That is definitely what we want to do. But as far as this particular contract is concerned, we tried to set it up in the best possible way, learning all that we knew about the past, to achieve that. It has proved particularly difficult for successive Governments to get the long-term unemployed into work at any price, right? And at the moment the emerging evidence is, this is below where it should be, which is why we are trying to do the performance, and we need to understand where people are making better results than others-I will not call them good again-to try to cascade them down.

There is an element-a small element-that says, "This is still one year’s data." You cannot get a job outcome until December 2011 for the first cohort, so we are talking about six months’ of data with another 12 months to go. We are one third of the way through, even of the earliest cohort, so a little bit of "Shall we just wait and see where this goes?" would not be inappropriate, albeit with good teams around.

Q59 Ian Swales: My final point on that, in terms of money saving or not, is what the obligations and expectations are around providers. A constituent of mine was going into a six-month job, or more-a permanent job-in the national park and he basically needed some kind of safety boots, or something. As I understood the way it works, it would have been in the interests of the Work programme provider to help him get ready for this job, but they did not. They said they had no money and he ended up going to some sort of charity to try and get these boots to prepare him for work.

There is something about being clear about what sort of help people are due to get and where the boundaries are, because my sense of it is that it is not providing the levels of support to people that were originally envisaged. Maybe that is because they are short of money.

Robert Devereux: When they set these contracts up-I am struggling to remember exactly how much money we said they had collectively raised. I think it was something in the order of £200 million. I am not sure quite where to go with that question, because if you have a contractor who is incentivised to get somebody into work and is prepared to receive the thousands of pounds that we are prepared to pay out and has not forked out for heavy duty boots, they are manifestly not optimising their position. At what point do I point out to them that that would be a foolish thing to do? I will be doing it with these performance teams, but it slightly calls into question the dynamism of the private sector if they cannot work out that a pair of boots might well be worth the investment.

Q60 Ian Swales: It does, and we always suffer from local management. I guess it goes back to the Chair’s question about parking and so on. It is the element of whether there are any standards that long-term unemployed people can expect from this process.

Robert Devereux: We had a long conversation about this the last time I came. We asked all providers to set down what their minimum standards were, and we did not set a minimum set of minimum standards. We said, "Look, you’re the ones running these contracts, so you’d better decide what standards you want to sign up to. If your standards turn out to be wholly unattractive and unworthwhile, and you’re not going to get jobs, the provider next door ought to do that." At the moment, as you can see, most of these providers are underneath, so there is a hard question about who is actually going to get to the right level.

At an earlier stage, I am afraid that I was interrupted on the point I was trying to make, which is that even within the 9.8%-in other words, the job outcomes for that first cohort-the range at the moment is between 5% and 14%. So the poor performers, even on the cohort at 5-quite a long way to go to 31. The 14 are halfway there, still with a year to go.

Q61 Ian Swales: So you are now talking about the providers by region. Is that the statistics? Or even nationally.

Robert Devereux: Across different providers. I have 40 contracts. On average, for that June cohort, they delivered 9.8% by July. The best one has delivered 14%.

Q62 Jackie Doyle-Price: You have just touched on what I wanted to raise. As time progresses-I completely get your point that we are on a cumulative journey here-you really are starting to see the poor providers from the good performers.

I want to address the point that the Chair made about parking difficult cases. The reality is that claimants on ESA are going to be the most difficult people to place in a job, quite often because they will have been claiming incapacity benefit for a number of years. They will have a number of psychological issues as well as physical ones. Would you say that it was realistic to expect their journey from entering the Work programme to finding a satisfactory job outcome of six months to take in excess of a year?

Robert Devereux: I think it is quite possible it would, yes.

Q63 Jackie Doyle-Price: My point is that I do not think we should be surprised at this stage in the journey of a five-year contract that those figures are so low. Did you anticipate that they were going to be significantly lower than the other groups?

Robert Devereux: I don’t know, because all I’ve got are the aggregate curves. If I had to guess, the answer would be yes.

David Clarke: Was it the same standards for all of the groups-the three groups that you set standards for?

Robert Devereux: Yes, but they do not include the hardest-

Q64 Chair: The important thing is, are they working with them? I don’t want to end up going by constituency, but the evidence from your own evaluation is that they are not. It might take longer, but they are not working with them.

Jackie Doyle-Price: I don’t think you can draw that inference from the period that we are looking at. It will come down to individual providers. If you look at the two providers-

Robert Devereux: You can both be right. It is perfectly possible that there is evidence of a parking agreement, which is what the research said. It is also perfectly possible that for those people who are being worked within these categories, it will take a long time to deliver an outcome. The reason for giving people two years is because if you said to yourself, "How do you fancy getting an ex-incapacity benefit ESA claimant into work?", the chances are it will take a long time. That is why we are offering to pay up to £13,700 for that outcome, which is of the order of four times as much as everybody else.

It is an interesting question whether or not these levels of payment are still too low to get people’s interest, but I am not sure that it will be in the Committee’s interest to suddenly ring up and say, "I’ve got a good idea. If I put the money up, will you do better?" That is not the right answer. The right answer is to say, "You cannot make money on this contract if you do not get people into work." That is exactly what the NAO established to start with, and consequently, if you put £250 million of your own money up, you are at risk of losing that, so your shareholders have got a very vested interest in making this go.

Q65 Guto Bebb: I want to clarify a couple of issues in relation to the structure of the contract, which means that you obviously have competition within areas. For example, in Wales, we have two prime contractors. One of the reasons for that was to allow competition, so that if one was performing better than the other, there could be movement in terms of the people referred between them. However, having spoken to a number of providers, my understanding is that that mechanism can only be applied if the individual contractors are actually hitting their minimum performance standards.

Robert Devereux: The first threshold is essentially that they need to be at minimum performance levels, because if they are below minimum performance, they are in jeopardy of losing the contract.

Q66 Guto Bebb: But at the moment nobody is hitting their minimum performance standard.

Robert Devereux: Correct, but also, this is a provision that we have made perfectly clear will apply from year 3 onwards, when we have actually seen two years’ worth of data.

Q67 Guto Bebb: My concern is very simple. The figures show a significant disparity between the performance of the best and poorest providers, yet nobody is actually hitting their minimum performance targets. Now, is there any mechanism in the contract to allow you to invoke that competitive element, even though nobody has actually hit the target?

Robert Devereux: Okay, I understand the question. I think I don’t know, but I will go and check whether the competition inside the market can still apply even if they are not at minimum performance. I am not currently aware whether, if they are not above minimum performance, the term does not apply, because that would be slightly odd-

Q68 Guto Bebb: That is my understanding, which is why I was asking for clarification.

Robert Devereux: Can I check and come back to you?

Guto Bebb: Yes, that’s absolutely fine.

Robert Devereux: Because that would be a very sensible thing to do, as far as I am concerned.

Q69 Justin Tomlinson: One defence you have been putting forward is the fact that it needs time to filter through the system. Looking at figure 4 on page 10, there is quite a variation between the top performer and the lowest performers. It talks about the number of contracts-is there any correlation with the time that those contracts were awarded? Is it the case that the lowest-performing ones are those that have had the most recent contracts, which would support your argument, or is there nothing we can draw from that?

Robert Devereux: I don’t think so. They did not all start at exactly the same time, but, broadly, June into early or mid-July is when they all started, so this far on I do not think that the precise date on which they started is important. However, I do think it is the case that we expected, certainly when we were doing some of the profiling, all of them to start pretty much with a zip when they took the contract. It is quite clear that some did start with a zip and some did not. I do not think it is the actual date; it is the efficacy of the work that they did in those early months that is probably the biggest discriminant here-whether they the actually put their foot on the gas to start with.

Q70 Meg Hillier: I have a couple of questions. First, on the employment support group, have you done or will you be doing any analysis of which bits of that group are doing better? For instance, my constituency has a high incidence of people with mental health problems that are on incapacity benefit. I have talked about the challenges they have getting into work, and we know that that is quite a difficult group to get into regular employment. Will you be disaggregating the outcome in your ongoing evaluation?

Robert Devereux: Again, I would have to check. That is not a particular data marker that would be part of the contract, because we do not ask that within the ESA. It would not surprise me if the research contracts we have got going, which uncover the sort of things that the Chair has talked about, indeed dig into that, but I will have to check.

Q71 Meg Hillier: Okay. If there is any information that you can give us on that, that is quite important, because the hardest-to-reach groups will be the most challenging.

Robert Devereux: You’re asking about whether there are mental health issues within each group.

Q72 Meg Hillier: Well, actually, yes. Sorry, I should have declared an interest at the very beginning. My husband is a non-executive director of a Work programme provider. Say G4S-there is no connection-was doing a particularly good job in its percentage on people with mental health problems, would you be learning lessons from that? You talked about learning lessons, but would you break it down to that level where there was particular expertise?

Robert Devereux: The answer would be, insofar as the research contracts that we have going are capable of looking at that level of detail, that yes, we would. I do not know what they are, so can I find out and let you know?

Q73 Meg Hillier: Okay. I am certain that we can ask those questions later. My other question was about the geographical links to performance and whether you have now analysed this list against geographical unemployment rates and so on, because in Teesside it is theoretically a bit harder, although in Hackney we have a 12% unemployment rate nevertheless.

Robert Devereux: But that does not seem to be an effect. In the last sentence of paragraph 15, the NAO concurs with our analysis that one of the possibly surprising things, which supports my view that this needs more time and more direct management action and not a change of course, is that this is not correlated with the labour market; it is to do with how well the provider is operating on their contract in their area. It is not driven by whether they are high unemployment areas or low unemployment areas. That is why I keep thinking that the right answer to this is to go back and confront these contractors with their low performance and ensure that they learn lessons.

Q74 Meg Hillier: In the level of detail that you are looking at this, are you doing any analysis? Will you be able to tell us at some point, if you cannot tell us now, whether there are any specific contracts between Work programme providers and different employers? We have heard before that some Work programme providers deal with all the vacancies for particular large supermarkets, for instance. Are you aware, or are there any rules about, whether money can change hands in that transaction? I am hearing that with one large supermarket chain-I have not had enough data to verify this, so I will not name the supermarket-there is a belief and some truth that it will only offer 16-hour-a-week contracts to people, which suggests a very direct relationship with some of the Work programme outcomes. Existing employees who need to up their hours, because of other changes that we were discussing earlier, are unable to do so. There is a big shared interest from the supermarket and from the Work programme provider in sticking to 16-hour-a-week contracts. Is that an intended outcome from the DWP? Do you care-do you mind-or is this something that is an unintended consequence?

Robert Devereux: If you invite the private sector to try to get difficult-to-help people into work, which we have all just established is quite difficult and it is finding it hard to do, and we say to them, "Look, you design your process, I am prepared to pay this for it", it is their choice what they then do with that payment as to whether or not they then change the market. I do not want to go around and say, "It is basically your choice to run this process as you like, but you may not do this and you may not do that." It seems to me a bit like the question of the boots: is it somehow market distorting for the provider to buy a set of safety boots for someone, and/or is it market distorting to come to some arrangement with a local supermarket about what hours it offers and what payment or-

Q75 Meg Hillier: I think it is a bit market distorting to say that it is easier to get part-time work at 16 hours a week in certain sectors than regular, full-time work. You gave that example of a woman with eight children. Full-time work with eight children is challenging-work with children is challenging, full stop-but actually if she was able to get full-time hours, she would have more income, anyone would, and that can ease your path in life generally-

Robert Devereux: But in that case she was working part time because of the children-

Meg Hillier: Yes, because of the children. Sorry, that was not a good example-

Robert Devereux: So the sort of job you just described is exactly the one she would need.

Q76 Meg Hillier: Yes. A lot of people do like those things, but if you want to work full time and that is the sector you are in, it is very difficult. The 16 hours seem to have become the golden mark.

May I give another example, Mr Devereux, although it is slightly off this particular subject? I had classroom assistants in my local primary schools being called in by their DWP advisers and told that they needed to leave their job and to work in the local supermarket to increase their hours from 10 or 11-typically, for a classroom assistant-to 16 in a supermarket. Many of them had got their life, finances and child care worked out around their hours in school. One of my heads, who is experienced, challenged the DWP to find the relevant rule, regulation, ministerial statement or whatever this was embedded in, but nothing was forthcoming, so that was then left. Another head felt pressured and somehow managed to manufacture an extra four hours a week for someone.

Robert Devereux: The current benefit system, as we all know, has got some odd rules in it. The whole point-not the whole point, as I keep coming back to, but an important point-is that within the universal credit world there are no artificial constraints on hours. We will let you work one, two, three, four or five hours, and for every single hour you earn you will get all the money for the first few hours, because that is what the disregard is all about, and after that for every single pound you earn you will get at least 65p-end of story. The thought that somehow or other people are adjusting the labour market to reflect how the benefit system is is obviously a daft way of doing it, and that is why we are changing it in universal credit.

Ian Swales: That is not new. That happened under the previous system as well.

Q77 Chair: May I ask you a final question before we wish you a happy Christmas? Why are you going back on your commitment to publish these outcome data every three months?

Robert Devereux: Because we did not commit in the first place. I think that there may have been a misreading of what is going on here, so forgive me. What we did say when we first started publishing the referral and attachment data was that it would be published quarterly. I am not aware that at that stage anyone said anything about the frequency with which the outcome data would be published. What we have now done is to publish the November figures-

Q78 Chair: November figures for what? Referral and attachment or-

Robert Devereux: No, no, sorry. What I am trying to explain is that there was an original pattern established of quarterly data on referrals and attachments. We have never said that we would do outcomes on a quarterly basis, so we have not changed our mind.

What we have said is that this is a brand-new series and the frequency with which the figures are published is a matter of judgment by the national statisticians who we have working for us. They determined that we have the first ones in November, although we wanted them earlier. We have also agreed and pre-announced that the next ones will be in May. That does not mean, though-let me be very clear-that we are saying it will be six-monthly. Like you, we all want to see the end-of-March 2013 data as soon as possible, so my expectation is that, having got this series stable-they well understand it, they know how to put the national statisticians’ flag on it-you will see the end-of-March data before the summer recess. So we will be back into much more frequent publication, at the point at which we know what the level of the series is-

Q79 Chair: We look forward to seeing you then. I hope that in future, in your press releases, you do not give credence to the trade data and that we stick to the outcome data, which I assume you okay. You must take responsibility for the accuracy of the information in your press releases, but all the data in there was unverified trade association data which I have always been very sceptical of, personally, all the time.

Robert Devereux: It was trade data.

Chair: It was trade data, and that should not appear in a Department press release-it appeared at the top of your press release, rather than the published and authorised figures from the national statisticians-

Robert Devereux: We have tried best to reflect what is a national statistic and what is not under our control but is none the less in the public domain and potentially relevant to the subject.

Chair: I hope that that is reflected in your forthcoming press releases on it. Thank you very much, happy Christmas and I hope that you will join us for a drink upstairs. Happy Christmas everyone, thank you.

Prepared 19th February 2013