Public Accounts Committee

Oral evidence: Sale of Student Loans, HC 1527

Monday 10 September 2018

Ordered by the House of Commons to be published on 10 September 2018.

Watch the meeting

Members present: Meg Hillier (Chair); Sir Geoffrey Clifton-Brown; Anne Marie Morris; Bridget Phillipson; Lee Rowley.

Sir Amyas Morse, Comptroller and Auditor General, Adrian Jenner, Director of Parliamentary Relations, National Audit Office, Simon Reason, Director, NAO, and Marius Gallaher, Alternate Treasury Officer of Accounts, HM Treasury, were in attendance.

Questions 1–119

Witnesses

I: Jonathan Slater, Permanent Secretary, Department for Education, Charles Roxburgh, Second Permanent Secretary, HM Treasury, and Justin Manson, Director, UK Government Investments.

Report by the Comptroller and Auditor General

The Sale of Student Loans (HC 1385)

Examination of witnesses

Witnesses: Jonathan Slater, Charles Roxburgh and Justin Manson.

Q1                Chair: Good afternoon and welcome to the Public Accounts Committee on Monday 10 September 2018. We are here today to look at the NAO’s Report on the sale of student loans. There is a good track record: the Treasury Committee produced a report covering some of this in February, and the Lords have published some work on it this year as well. We are looking particularly at the sale of the first tranche of loans that are income-contingent. The first sale in the series aims to raise about £12 billion by 2022. It is worth saying, because there has been a lot of confusing publicity about this, that the selling of the loans has no impact on the 410,000 borrowers who were on Mr Slater’s books at the Department for Education, but whose loans have now been sold on to other investors. The key thing we want to touch on is that the sale enabled the Government to reduce public sector net debt by £1.7 billion, but although that was a win for today, it forgoes about £3 billion of future income, which has an impact, of course, on the Department for Education, although we know that the Treasury has worked to resolve that.

We want to explore some of the technical stuff today, but the NAO has concluded that the way the sale was handled and the market worked was pretty good, so we want to talk about the real-world impact and the methodology and approach taken by the Treasury, as well as the fact that the Government have a policy to sell anything that they do not need to keep for policy reasons. We want to explore what the policy reasons might be for doing or not doing this, bearing in mind that coming down the line are other loan books, some of which are quite different in nature.

Perhaps I can start with you, Mr Roxburgh. Sorry, I should introduce our witnesses first. Charles Roxburgh is second permanent secretary at Her Majesty’s Treasury, and Jonathan Slater is permanent secretary at the Department for Education—both of you really need no introduction in this Committee. Justin Manson is the director of UK Government Investments. Welcome to you.

Mr Roxburgh, given what I have said about the Government policy, why did you decide particularly to sell these student loans?

Charles Roxburgh: As you say, the overall Government policy is that we should not hold assets for which there is no public policy reason to hold them, and provided that we can sell them in a way that raises value for money, that meets our value for money tests. The public policy reason for extending loans to students is to fund their higher education. Once they have graduated and entered the workforce, there is no public policy reason to continue to hold that asset, and therefore if we can sell it in a way that realises value for money, it would be our policy to do so.

Q2                Chair: Except that of course the money the students are paying back on the loans could be revenue to Government to re-lend out to future students. Was that something that you considered?

Charles Roxburgh: That is absolutely central to the estimate of value for money. The way we assess value for money with these loans is that there are three tests. There is whether there is an efficient market and whether we can get sufficient pricing tension, to get a good price for them, but the critical quantitative test is whether we can sell the loans for more than the retention value—the value if we retained them, held them on our balance sheet, and gathered those future cash flows in. That is why what is critically important here is to model that retention value, and our colleagues in UKGI developed a very sophisticated model to do that, which the NAO reviewed, and it passed muster; there was a strong vote of confidence in that model. The key is to sort of discount those cash flows to see whether, if we retained them, it would be worth more in present value terms than if we sold them. That was the essence of the assessment, so the fact that we sold them for more than the retention value is why we are confident this met the value for money test.

Q3                Chair: We are hampered—well, the NAO is not hampered, but the rest of us are, because the retention value range is obviously not a public figure. That is the case for reasons that are understandable, but it is frustrating, as I think we would all agree. So Mr Slater, quite a complicated model was set up to work out the value of all this; in more human terms, what has been the impact on your Department of the sale of this loan book?

Jonathan Slater: In human terms, I guess a number of us have become a lot more expert on selling the student loan book than we thought we would be when we joined the Department for Education. That wasn’t among the list of criteria when they chose me. Luckily, I of course have some experts in the Department, and it’s why, of course, we work so closely with UKGI and the Treasury. I am the accounting officer—I own the student loan book—and I have got to tell the Secretary of State that it is value for money, in accordance with Government policy, to crack on, so I have spent a long time getting to understand the Green Book and securitisation in detail so I could make that judgment.

Q4                Chair: Was it a partnership of equals? You have highlighted that you are not the only permanent secretary who didn’t come in with expertise and knowledge about selling off these financial vehicles. It is Government policy, and the Treasury drives some of that. Did you feel that you had the power to really look at this and to say, “No,” if you felt it wasn’t value for money, or was it Treasury-driven?

Jonathan Slater: Well, all decisions are decisions that the Government makes, not civil servants. My job is simply to advise the Secretary of State whether it is value for money or not. If I thought it was not value for money, as determined by the Government’s approach to value for money in the Green Book, I would have been very happy to say so, and he would have decided whatever he decided. I thought very carefully about it, and I spent a long time making sure I understood in detail the proposition in front of me so I was able to say to the Secretary of State that it passed that test, as the Government defines value for money. Obviously, I am delighted that the NAO agree with me. If I thought otherwise, I would have said so.

Q5                Chair: The face value and the carrying value in your accounts were different, but then the selling price was substantially lower than the carrying value in your accounts. Do you want to explain that to us in layman’s terms?

Jonathan Slater: Yes. I am much better at explaining it in layperson’s terms. If there are any differences between my layperson’s account and Charles’s technical account, go with him, but I am going to do my best to explain in simple terms what the difference is.

When it comes to valuing the loan book in our departmental accounts, what we are doing is trying to work out today what a future income stream is worth. The first thing we have to do is to work out how much money we would expect to get in the future, which obviously isn’t the face value because as you know we are not expected to get all the money back from graduates. It is designed so we won’t get it all back; we will only get a certain proportion back, depending on how much money people earn. The first thing we do is discount the face value by the amount that we don’t expect to get back. That is straightforward.

Then for a transaction of this sort, we have to follow international accounting standards for how you would account for it. The simple way I think of it is that a future income stream tomorrow is the same as borrowing some money today and repaying it with the money you could be getting back in the future. That is what the value today is of a future income stream. To work out that sum of money, having first discounted it by the amount of money that we are not going to get back, we have to apply an interest rate. The Treasury tells us what the interest rate should be, based on the international accounting standards. It is 0.7% plus RPI. That is the Government’s long-term cost of borrowing. It is not the exact cost of borrowing today, because obviously if we used today’s cost of borrowing you would revalue the student loan book every day.

That is how we do the departmental accounts, and it seems to me perfectly plausible. Those are international accountings standards, so there you go. That seems to be a perfectly sensible way of doing it. That’s not what you are doing when you are selling it. You are doing something different when you are selling it, in three ways. Sorry—you sometimes accuse me of giving lengthy answers—

Chair: No, I think it is helpful to set out the baseline in layman’s terms.

Jonathan Slater: This is essentially my core contribution to this meeting. There are three reasons why the sale price is different. First, we are not selling the whole student loan book; we are selling a particular element of it. We are selling the first loans—the £1,000 loans—so the amount of money that we expect to get back from those graduates is not the same as the average for the student loan book as a whole. It is a piece, and you have to work out what you think the particular amount of money you are going to get back from those people is going to be, which is why we designed a model for that purpose—the NAO was comfortable with it. That is one reason why the figure is different—because it is a specific loan book, not the whole thing.

The second reason is that if you are selling something, what we are getting is guaranteed cash now instead of a contingent income stream down the track. We are essentially transferring the risk that the money comes in from us to the sellers, so they are obviously going to ask for a payment in return for us transferring that risk to them. That is self-evident. That is the asset-specific risk.

Thirdly and finally, when we are working out what Charles rightly calls the retention value—what it is worth to keep the thing—the alternative, in this context as opposed to accounting, is not borrowing more. That is not an alternative at this point because the way the Government does it is to set its overall fiscal strategy on the basis of a certain level of borrowing, as you know; it sets its targets for borrowing. If we were not to sell the student loan book, we would not have the opportunity to borrow more. What would we be doing? We would be spending less. The whole point of selling the student loan book is to enable you to spend things without increasing debt or taxing more. That is the whole point of it.

When we consider the alternative—this is the last bit, I promise—what would it be? It would not be borrowing more, so we do not use a borrowing interest rate. The alternative would be not spending the money. What we do to calculate the retention value is to say, “Okay, if we do get the money in from the student loan sale, we could spend it on roads, schools, hospitals or whatever, and they would generate a rate of return for us. How much of a rate of return? It would be 2.5% plus RPI, because that is the amount the Treasury requires as a rate of return before it will let me spend anything on a school.” That is the calculation I am doing. If I get the money now, Government can spend it on something with at least 2.5% plus RPI, so as long as the discount I am paying is not more than that, it is value for money.

It all rests, critically, upon the fact that Government cannot increase its level of borrowing. It has set a policy framework that sets it as it is. That is why it is different from borrowing the way you do in the accounts, because that allows borrowing. That is my best attempt to explain it.

Chair: At this point I will transfer to Mr Rowley. I have some things I will pick up on, but Mr Rowley can go first.

Q6                Lee Rowley: How would you make a layman sitting in my constituency today understand the five-minute explanation you have just given and accept that the Government has done the right thing here, having exchanged ostensibly £2.5 billion to £3.5 billion of money for £1.7 billion?

Jonathan Slater: Well, I would not rate my chances highly, but the one-line sentence, as Charles has said, is that we have sold it for more than we would get if we kept it. That is the one-line answer to the question. We will not get £3.5 billion, so the £1.7 billion versus £3.5 billion—

Q7                Lee Rowley: You might get the £2.6 billion, though.

Jonathan Slater: We might get that, but we might not. The advantage of having definite money today versus possible money tomorrow might persuade a layperson that it is worth something. If I offered you the possibility of £10 tomorrow or definitely £9 today, that is something a layperson can get their head around. The final thing I would say is that how much cash you get does not necessarily matter so much as what you do with it—right? If you get that £1.7 billion and you invest it in things that generate more and more value, so that the value they generate is more than the money you would get in the future, that has to be worth doing. I might get out of the pub with an agreement.

Q8                Lee Rowley: If this whole process of the £9 versus the £10 is so inherently risky, why is the Government in this business full stop? Why does it not just encourage the private sector to construct a process that would allow these kinds of loans to be given to them and the entire payback, whatever it is, to be done in the private sphere?

Jonathan Slater: Are you talking about a regime in which the private sector collected the money?

Lee Rowley: I am not being specific. I am harking back to your point, which is, “This is very risky and we need to get them off the book.” If that is the case, why are they on the book in the first place?

Jonathan Slater: I will start and then Charles can come in too. The starting proposition is that, as the Chair said at the beginning, Government policy was that borrowers would not be impacted. That was the basis on which the Sale of Student Loans Act 2008 was passed by Parliament. We are not in the business of transferring to the private sector the ability to collect money back from the individuals themselves, if they can do better. That is not the deal. The policy says it should not make any difference to you, as a borrower, whether the loan has been sold or not. That is the starting point.

Revenue and Customs collects 96% of the money through the tax regime, so it is not as though there were a tremendous opportunity here for private sector innovation anyway. The money will be coming in to us, nearly all of it through the tax regime. The question is whether we can get a value for money proposition by asking the private sector to give us money now in return for that cash flow in the future. For the reasons that I have tried to explain, so long as you accept the Green Book and Government policy on value for money, it passed the test, as the NAO have said.

Q9                Lee Rowley: The whole premise you are explaining is on the basis of reducing the risk to the Government, right? So as of a few months ago there was £102 billion of student loans on the Government’s books. Given that, according to your assessment, it is so inherently risky, what is your plan to get rid of it as quickly as possible?

Jonathan Slater: We are not going quite that far. What we are saying is that the question for civil servants, as we advise Ministers, is whether selling something today generates more money that retaining it. If it does, we recommend selling it, and if it does not, we do not. We are saying that one of the variables that you use to make that judgment is the risk that you do not get the money in the future. The other variable is the benefit of having money today to invest in things versus having it in the future to invest in things. We look at it dispassionately and Government has decided that it wants to generate about £12 billion of receipts by 2022 of plan 1 sales—it has not made any decisions on plan 2—so we will crack on with that programme and, so long as the market keeps offering us more than retention value we will recommend selling. If that were to change, we would recommend keeping.

Q10            Lee Rowley: All of this assessment is based on your models, as good or as bad as they can be and as assessable as they are at this stage. We accept that the model has variants in it. The only thing you can say with certainty about the model is that it will not be right, as models are never right—it is quite close to being right. We accept all of this—the value for money framework is being met—on the basis of a highly complex model. At what point do we get to where we know whether that model has worked or not? How do I as a politician know whether the extreme expertise that is being applied is actually correct?

Jonathan Slater: I suppose you can take some initial comfort from the NAO’s review of the subject matter. Equally, obviously we will keep things under review as we go and adjust as we go. One of the things that the NAO draws attention to is that the model that UKGI had to build, designed to value the specific loans we are selling as opposed to the loan book as a whole, has a 5% tolerance in it, precisely for the reason you described, and—good news—it is remaining within that 5% tolerance. Clearly, if things were to change, we would adjust our models accordingly.

Q11            Lee Rowley: We can adjust for future sales, not for this sale, as this money is gone—other than the profit share, which may or may not be applicable. We are not going to get anything more back than the £1.7 billion. So we are taking it that your model was correct and that we have not got the wrong amount of money coming in to us—above or below what it should be. I know I am getting into quite theoretical points about models, but actually your entire thesis here is based on modelling, and it is modelling that we cannot prove or disprove at this stage.

Justin Manson: Your point is absolutely right. It is impossible to know for certain what the outcome will be until the very end of the period, so we take that for granted. So given that, what we try to do with the model is put in our absolutely best resources and evaluate as many ways as we can. So not only have the Government Actuary’s Department statisticians and operational research people been involved in this; it has also been reviewed by our advisers, it has gone to the rating agencies and investors have looked at it and examined it. So it has undergone a very high degree of scrutiny.

Over time, as the data points come in, we will know how it is performing, but as you will know we get essentially one data point a year to be able to judge that. As Mr Slater said, so far we are quite within the bounds of tolerance, but I do not want to make any predictions about where it will be next year or in 20 years’ time beyond what we have put in the model already.

Q12            Lee Rowley: As the accountable officer, Mr Slater, if the Treasury had not come to you and said, “We want to look at selling these off”, would you have decided to look at this independently?

Jonathan Slater: No, that is not the way that it works. The Government have an overall fiscal strategy, in which the Treasury work out the potential for the student loan book to generate some value. They then come to me—well, it was years ago; they came to our predecessors in BIS, as it was—and says, “Here is a proposition”, and we then work together on whether it will generate value for money.

I am the accounting officer, because I own the student loan book—for reasons we can discuss if you would find that helpful—so I am the one who has to say to the Secretary of State whether it is value for money. I don’t get the money to spend; it is Government money, set in the context of an overall fiscal strategy. That is why it starts with the Treasury.

Q13            Lee Rowley: I accept the point about the Treasury commencing this, but I am asking you, as the accounting officer—you are the closest person to this—whether you would have instigated such a policy, given your knowledge of these issues?

Jonathan Slater: I think that overstates the importance of the role of the permanent secretary of the Department for Education. Parliament decided in 2008 to introduce an Act to sell student debt. Parliament made that decision then because they saw the potential for precisely the sort of proposition we are discussing today.

Q14            Chair: Mr Rowley’s question is quite simple, really. Aside from that—we are maybe moving into theoreticals a little—you are the accounting officer for the Department, which had a bird in the hand of £1.7 billion, as opposed to jam tomorrow. I am mixing my metaphors rather badly, but you know what I mean. Would that have been uppermost in your mind, with all the other things you are doing in the Department? Would you have thought that it was a good thing to suggest if it hadn’t been wider Government policy? Would it be something that you might have suggested in the current financial climate, because of the need to manage the budget, to your Secretary of State? Do you think that is something that would’ve occurred to you? You acknowledged earlier that you didn’t have the financial qualifications.

Jonathan Slater: No. That is not the way that the civil service is set up. It doesn’t employ permanent secretaries in the Department for Education to come up with ways of increasing assets for Government to spend on things. That is typically what the Treasury does. They then come to me and I have to judge whether it is value for money. That is the way that the process works; you wouldn’t expect Charles to be thinking of clever ways of improving school standards. However, if I needed his help, I would come to him, and he would make a judgment and so on. It is that way around. I don’t think there is anything surprising about that.

Q15            Chair: On the modelling, Mr Manson was involved and has a background in the private sector. We are interested in why the Government did the modelling, rather than the private sector. You must have crawled over that, to make sure that, given your previous experience, you weren’t selling something too cheaply to the private sector. Do you want to talk us through what checks and balances you personally oversaw to make sure that that was right? We talk about the Government acting responsibly. There were a lot of people involved in this. How can you reassure the taxpayer that this wasn’t given something away too cheaply?

Justin Manson: There are two elements to it. One is the model itself and the other is the market testing. There are two elements to the exercise. One is the work that we did on the model. It is my view that, whenever anybody looks at selling an asset, wherever they are, they need to do their own work and come up with their own point of view on the value. We did that using the absolute best set of resources that we could and getting the scrutiny that we could from a wide variety of interested parties.

The second piece is obviously it being tested in the marketplace, which goes back to the comments that Mr Roxburgh made about an efficient market and an efficient price. It was clear to us relatively early on that a securitisation format gave us the best prospects of involving as wide a group of investors as possible to bid for the securities, so that, if we had a competitive bid for each of the securities, there was an additional measure of comfort that we were achieving a price that represented good value to the taxpayer.

Q16            Chair: Going back to your previous experience, would the private sector have done a better job at all of developing a model? We can understand, from what the NAO said, why that didn’t happen, but do you think you would have had a range of different models, or do you think they would have come to about the same place? Obviously they didn’t have access to the HMRC information—that is one of the reasons the Government did it—but do you think there would have been a material difference if the market had developed its own model, compared with the Government doing the modelling?

Justin Manson: No, I don’t, actually. I think the private sector would have found it quite difficult to replicate the kind of model that we did. It was an extraordinarily time-consuming process with a huge number of data points. There are a variety of reasons why the private sector would have found it difficult to replicate it. We talked about having private sector players develop a separate model and we got reasonably far in doing that, but we concluded that the model we developed in-house tested much better than the one we talked about with the private sector.

Q17            Chair: Will the work that has been done on that model be useful for future loan book sales?

Justin Manson: Yes, it will. You are right that the model we used was specific to the cohorts of loans sold, but the basic techniques that we employed are replicable across different cohorts. Needless to say, adjustments will be required for the different characteristics.

Q18            Chair: Even for the ones that are currently being lent, which will end up potentially being £50,000 over a period of time?

Justin Manson: We have not looked at those loans. I was answering you in the context of the plan 1 loans, not the plan 2 loans. We have not developed a model for plan 2.

Q19            Chair: At different times, different politicians have discussed changing student loan models. One potential change is raising the income threshold at which it is payed back. To be clear, is this particular sale locked in at £18,000? [Interruption.] It is locked in? Forgive me, I forgot. I think Sir Geoffrey has just answered my question.

Jonathan Slater: The Government have made clear that they are carrying out a post-2018 review not of plan 1 loans, but of plan 2 loans. Clearly, there is nothing to prevent a future Government making the changes they want to make.

Chair: Then you’d have to pay compensation.

Jonathan Slater: Then we’d have to compensate them, which is why the Government have made it clear that they have no intention of changing the threshold or any of the other terms for such loans.

Q20            Sir Geoffrey Clifton-Brown: Can I come in on the model? Paragraph 3.11 shows the profit-sharing numbers. What work did you do to build those numbers? It seems that they are so high that there will never be a profit-sharing agreement.

Justin Manson: There are two approaches. One is trying to develop a sophisticated model for this. More realistically for this particular element of it, we based it on discussions with investors for that specific tranche of notes, where there is, as you know, a much riskier set of pay-offs than the notes that are higher up the capital structure. The numbers that we came at were based on how far we could push investors to take, for example, the length of the profit share buy-back, the multiple of the expected return and so on. As a result of that dialogue, in my judgment and other people’s judgment, we came to the best compromise that would solve that.

Q21            Sir Geoffrey Clifton-Brown: Having been through this process and looking forward to the next tranche of sell-offs, are there any particular lessons that you have learned and things that you would do differently next time?

Justin Manson: The NAO obviously has some recommendations, which we agree with and have begun incorporating. The other piece—perhaps this is more responsive to your question—is that we have gone back to investors, some of whom did not buy the first tranche, to understand their feedback. We have talked to people who have already invested and asked for their feedback. We have been doing that over the course of the summer with around 20 names to refine our judgments about things such as the securitisation structure, what kind of tranches we should have and whether we should change them, whether there are specific things in the way that some of the pricing is linked to different metrics that should change, and whether there is a way that we can be responsive to new regulations about what kind of investors can buy them. We are thinking about all those things, some of which are lessons learned; some show us responding to changes.

Q22            Sir Geoffrey Clifton-Brown: Finally, paragraph 2.14 says that the novelty premium and asset premium are about 4.9%. In the sales of future tranches of loans, would you expect the novelty premium to reduce?

Justin Manson: Yes, I would expect it to reduce. I hesitate to give you a quantification of how that might reduce in the future, and it will be quite hard to measure whether it was the novelty premium that went down or whether it was some other factor, but we absolutely would expect it to decline.

Chair: Would there be any benefit in delaying? We are thinking the same thing; carry on, Sir Geoffrey.

Q23            Sir Geoffrey Clifton-Brown: In terms of the timing of the next sale, it clearly takes time for this whole thing to bed in and for existing investors to feel comfortable with what they have done. Is this having an influence on the timing of future sales?

Justin Manson: It will, in the sense that we will be looking for indications of where this is going to price before we make a recommendation to the Department or to Ministers on whether they should proceed. Some of that will have to do with what feedback we are getting around investor comfort on some of the kinds of issues that we would have ascribed to the novelty in the first tranche. To be clear, there are a number of elements in that novelty premium that will take a very long time to work their way through.

There are two points on that. The first is that we get the data point once a year for the securities. In other securities markets, you might have real-time continuous updates and information, or it might come quarterly or whatever, but this is annual. It is important to bear that in mind. The fact that the model has performed well to date, well within the Department for Education’s threshold of tolerance, suggests that it is well accepted, certainly by us and by investors, as performing reasonably accurately so far. That is supportive.

The other element that is relevant is the liquidity of these securities. These securities will never be particularly liquid, because they are very niche instruments for which direct comparables in the marketplace are not available yet. Hopefully, with a programme, there would be some comparables over time. We will have one comparable, which will be this one. That will help to address it. But, again, because it is relatively illiquid, it does not give you quite the same accuracy of price information that you would have with something very liquid. Maybe that is too long an answer to your question.

Chair: That is very helpful, because we are keen to look at any lessons for the future.

Q24            Anne Marie Morris: Mr Manson, I suspect you had to deal with some conflict between the two Departments, because they have slightly different objectives to achieve. You are clearly looking at it from the market perspective, but the two Departments measured in slightly different ways, and we did not use PSNFL. When you gave your advice, how did you manage the conflict in what they were trying to achieve?

Justin Manson: I have to say that I did not see a lot of conflict, so I will start with that, but more importantly, when we look at the objectives, they were very clear. You may have a point in your mind about whether some other objective or some other metric would have been more appropriate, but in terms of what UKGI was working to, those were really very clear.

Q25            Anne Marie Morris: Whose was the final decision, then? Was it yours? How did you make the decision between the three of you as to what that final model would be?

Justin Manson: Sorry, the final—?

Anne Marie Morris: You have three Government Departments and you have one model. Who finally decided that that was the right model?

Jonathan Slater: Actually, we did not have one model. As the NAO Report identifies, we had the retention value, and we were using four different valuations and cross-checking them against each other as part of ensuring that we minimised the risk of getting it wrong. The decision maker on the sale price is the Secretary of State for Education, advised by me as the accounting officer on whether it achieves value for money. I take advice from UKGI, which is operating with a rulebook set by the Treasury. It is not the case that we had different objectives. The Government set a set of objectives. He has a rulebook and some experts, and I bring that together: “Does it meet the value for money test for the accounting officer? Yes,” and the Secretary of State sells. That is the process.

Anne Marie Morris: Okay. So a conceptual project that you would never have set out on—I think that was what you said in your reply to Mrs Hillier—was the project on which the Secretary of State was the final determinant. I find that interesting.

Jonathan Slater: It is not that I would not have done it if it were me; if that was the question I was being asked, then I gave a misleading answer. The question I thought I was being asked was whether, in a theoretical world, I would have come up with that idea myself. The answer is no, because that is not my area of expertise. That is all I was saying. The Treasury come up with the proposition; I have to consider whether it is value for money and I conclude that it is, as defined by the Green Book, rather than my coming up with my own definition of value for money, which obviously I am not allowed to do.

Q26            Anne Marie Morris: Okay. Mr Roxburgh, I think you wanted to say something.

Charles Roxburgh: As Mr Slater said, the Chancellor and the Government of the day set the overall fiscal framework. Asset sales, across a whole range of assets, have a very important part to play in that, provided that we can sell them in a way that meets the value for money tests, brings down debt and allows the Government to invest in other important public services and public investment that fulfils other goals. The Chancellor sets the overall fiscal framework. We talk to all sorts of Departments to see whether we can find opportunities to sell assets in a way that realises value for money and frees up capacity for more productive investments.

Q27            Anne Marie Morris: Mr Manson, as this was a bit of a one-off and there had not been a sale of anything like this before, you adopted the securitisation model, which I understand, but how did you persuade investors, given that this had never been done before, that it was something worth investing in? What was your case?

Justin Manson: There are a number of unique features of this securitisation. One of them is that it is inflation-linked. Another is that the returns are driven by increases in real wages. There are a number of macroeconomic variables for which it is actually quite hard to find other instruments where you can make those kinds of plays. That is a reason to be interested. Another is that the programme is likely to be of a size sufficient that investors think it is worth spending their time to understand what is a very complex model by most securities standards. Then it is down to the question of value. You will have seen from the NAO Report that we went to about 200 investors, of which roughly 60 eventually invested.

Q28            Anne Marie Morris: Who were these investors?

Chair: Figure 15 on page 35 gives some numbers on the investors.

Justin Manson: Perhaps I can give you a sense of what kinds of investors they were. There were pension funds and insurance companies. Those were expected to be the bulk of the investors in terms of proceeds, and indeed that is what happened. Other investors included banks, banks’ treasuries, private wealth managers, alternative asset managers, hedge funds—a relatively broad number of investors.

Q29            Anne Marie Morris: Are their names public?

Justin Manson: No, the names are not public. We have not made them public because, at least in our view, it is very commercially sensitive information.

Q30            Anne Marie Morris: In what way is it commercially sensitive?

Justin Manson: Because we would expect to go back to other investors in future programmes. Some of these investors do not want their names disclosed, and we do not think it would be helpful to the success of future sales for those names to be public. If it is helpful to you, we are happy to write in commercial confidence.

Anne Marie Morris: Okay—that would be helpful.

Q31            Chair: We will take that, but why are they so reluctant? They are taking on student loans, which were funded by public money in the first place, so they are playing a role in public policy. Why are they so squeamish about being named?

Justin Manson: I am not saying that is true of every investor—actually, it is probably not typical—but having agreed not to disclose the names of one or two investors, we obviously do not do it for the rest of them.

Q32            Anne Marie Morris: Given the identity of those who actually made a decision to go ahead and invest, and the larger pool of interest you had right at the start, when we go for the second tranche, what is your sense of who among that group—which of the different categories you identified—will be most interested? Did you find after you had talked it through that it was clearly of more interest to a particular group, and is that where you are going to go next time?

Justin Manson: I expect that some of the investors who did not participate in the first round remain very interested in the security as a concept, but it will depend on value. There are a number of investors that got all the way to the finish line but could not agree to the price we wanted to set. I would expect those to come in this time.

Q33            Anne Marie Morris: But none the less you had enough interested investors that you had no problem satisfying at a price you were happy with.

Justin Manson: No, not at all. We had over-subscription in each tranche. We would expect there to be some investors who invested in sale one who may not come back; but there will be others that did not invest who will come in.

Q34            Chair: Which of you agreed not to name the investors? Whose decision was that?

Justin Manson: That would have been advice that we would have put to the Department, obviously, working alongside the Treasury.

Q35            Lee Rowley: Explain to me the commercial sensitivity of being named publicly as buying, effectively, Government debt.

Justin Manson: It is a characteristic of some investors that they do not actually disclose the investments that they make, and they did not particularly want this to be the first time.

Q36            Lee Rowley: So they set terms, and the rules on the playing field, then, did they? If they don’t want to do it, we don’t do it.

Justin Manson: Well, I think our emphasis here is on achieving value for money, so the extent to which we saw real, genuine interest from investors in the security was the paramount thing in our mind, rather than disclosure of particular names.

Q37            Lee Rowley: You would accept there is a transparency issue here.

Justin Manson: Of course. Any time you do not disclose something, there is a transparency issue.

Q38            Lee Rowley: So when does the transparency issue become too much to be acceptable? What is your line?

Jonathan Slater: Again, these are all decisions that Ministers make on the basis of advice from officials in the normal sort of way, and the advice from people who do this thing for a living is you get a better price if you apply the confidentiality rule than if you do not. That is a choice, obviously, that can be made—

Q39            Chair: Can we just be clear, though? Mr Manson, maybe I am mishearing you; did you say a couple did not want to be named?

Justin Manson: I do not want you to hold me to a couple, but my point was that it was very few.

Q40            Chair: Yet if you look at figure 15, the number of investors placing orders was quite significant, so even if a handful did not want to be named, was that really going to have a big impact on the commercial price you were going to get? This is obviously advice you gave Ministers, but in this case I would ask you to share, broadly, your thinking on that with us, because it helps us understand.

Justin Manson: At the time we would have made that decision, we would not have known exactly who was going to be in the book, so we would have taken interest from very specific committed, serious investors that had done work, and we would have taken on board a comment that they would have made about whether they wanted their name disclosed or not. The ultimate composition of the book—

Q41            Chair: Sorry, you say “a comment they might have made”; was that a request not to be named, or was it just—

Justin Manson: It was a very direct request.

Q42            Chair: Why did they not want to be named? You said they do not always want to reveal their books, but if these people were investing my private money, I would know if I was in that fund; I would have a report from them, wouldn’t I, about what they were investing? So it is not secret what their investments are. There are people who know that their pension money or their funds are being invested. So it is not a secret; it is just not in the public domain in this context.

Justin Manson: Well, you are absolutely right about some. Some funds disclose what they are invested in to their investors. There are other funds that do not always disclose that.

Q43            Chair: So why did they not want to be named? Can you just explain?

Justin Manson: My understanding of it is they did not want to be disclosed because they are not typically disclosed.

Q44            Sir Geoffrey Clifton-Brown: I suppose some of them may well be in the category you mentioned of private wealth managers, who may well not want this to be disclosed. I suppose the acid test of whether you think you made the right decision is whether you will make the same decision when it comes to the sale of the next tranche.

Justin Manson: UKGI’s advice to the Department for Education and the Treasury would be to follow the exact same procedure.

Q45            Lee Rowley: What assessment have you made of the gain that you achieved on the basis of lack of transparency, financially?

Justin Manson: That is a very, very difficult question to answer.

Q46            Lee Rowley: You are in the realm of complex modelling, by your own account.

Chair: And now you have got the number of investors, so now you know who invested in the book. You have got them all—

Justin Manson: The facts will change for the next sale. It may be the same investors; it may be different investors. They may see—

Q47            Lee Rowley: The sun might not come up tomorrow. There are lots of things that might change, but fundamentally you have a transaction; you should be able to model what you think the gain was, in terms of the lack of transparency. Have you done that?

Justin Manson: No, I have not done that.

Q48            Lee Rowley: So how can you give good advice if you do not know what the model or the assessment is?

Justin Manson: Because the assessment at the time was that the judgment is extremely subjective, and a marginal judgment, and we were far from having perfect information about who would end up in the book at prices that we felt satisfied with.

Jonathan Slater: It is a legitimate question. We discussed earlier what we will do differently next time round. We have access to information now that we didn’t have then, and as Justin said, we will take advantage of that next time round. He is being honest with you that he is not confident in his ability to actually put a number to the extent to which transparency would have made a difference in the price. Equally, it is my job to make a recommendation to the Secretary of State on value for money, so I will test him on the extent to which it is possible to answer that question.

Q49            Chair: But your job, as permanent secretary, is to make recommendations to the Minister about other wider political or policy issues. Transparency and openness in government is something that Government have been trumpeting as important for some time.

Jonathan Slater: I am offering you something here, Chair; it may not have become clear. The first time round, as Justin explained, he had to make that commitment in advance of knowing how many bidders he was going to get.

Q50            Chair: So you think it might be different next time?

Jonathan Slater: The second time round, we will have some information. Justin is being completely straightforward with you: he doesn’t think it is possible to come up with a quantifiable number against transparency of the sort that I and the Secretary of State would obviously like, even with the information at his disposal now. However, I will press him on that point before we get to the second sale.

Chair: We are speaking for taxpayers and consumers of services—people who have taken out these loans. There is always suspicion when names are in secret. On private wealth managers, the names of the people they invest for don’t need to be public—that is the whole point of private wealth management—but making public the names of whichever private wealth fund is investing doesn’t seem to me to have any impact. Knowing what we do now, particularly with hindsight—I get you on hindsight—there are plenty of people in the list of investors in figure 15, so if a couple drop out because they are squeamish about being named, would that really have an impact? That is what you are debating, presumably.

Jonathan Slater: Lee Rowley is asking whether we will be able to get a figure next time that lets us make a judgment as to whether transparency is worth the price. It is a perfectly legitimate question, and one I would expect my Secretary of State to ask me. I will do my best to get whatever information I can from Justin. He is being honest with you that he doesn’t think he can do very well, because of the difficulty in being precise. However, we will go through that process in a way that we couldn’t last time, and you will hold us to account on the second sale.

Q51            Chair: We are going to move on from modelling. Mr Slater, you talked earlier about the endorsement—I think “strong vote of confidence” was the phrase you used—of the model given by the NAO. In paragraph 3.18 on page 34 in part 3 of the Report, the NAO actually says that the model is “within tolerance levels but it highlights the uncertainty”. That’s not really a ringing endorsement, and “strong vote of confidence” is probably a bit strong, too. Do you want to slightly recalibrate what you said earlier?

Jonathan Slater: I suppose I am used to sitting in front of you without a value for money judgment from the NAO, so when I get one, I rejoice. I wasn’t making a specific—

Q52            Chair: So this was a ringing endorsement, in Education terms?

Jonathan Slater: I didn’t use the phrase “ringing endorsement”, and I wasn’t making a specific point about that particular model. I was saying that if you accept the Government’s approach to assessing value for money, in accordance with the Green Book, Amyas says it is value for money. That is all I’m saying. I don’t normally get that from you, so that is all I’m saying.

Chair: It is fair to say that I acknowledged at the beginning that there are aspects of this that the NAO concluded had run well. We were keen to explore some of the wider issues—on which point I will ask Mr Rowley to come back in.

Q53            Lee Rowley: The Financial Times doesn’t think it is value for money. Why should we believe that it is?

Charles Roxburgh: I assume you are referring to the article that referenced the difference between the sale price and the Departmental accounting value.

Lee Rowley: There were several articles.

Chair: Are you referring to the article on 7 September?

Charles Roxburgh: That goes back to the explanation that Mr Slater gave, which is that the value of these assets held on the Departmental accounting balance sheet is not formulated in the way that the Government value assets for sale. That is because we look at the risk of holding those assets, which is not reflected in that discount rate, and the opportunity costs, because, under our approach to asset sales, we consider the cost of having capital tied up in these assets when it could be used to build other products and services that would develop benefits for society. We look at the value of these assets differently when considering whether to invest or sell assets from the value they hold on the Department’s accounts.

Q54            Lee Rowley: This is a sale of £3.5 billion of loans, on book value—on what is in the accounts. Would that on its own have an effect from an interest rates perspective on the PSND reduction? It is non-material, right? It is a drop in the ocean compared with a lot of other problems we have in the public finances.

Charles Roxburgh: No, I think the overall student loan programme is material. It is £12 billion over the forecast period; that is a material element in the fiscal strategy, provided that we can do it in a way that is value for money. Coming back to what you said earlier about simple explanations, as the NAO Report said, these loans originated in 2002 to 2006. Half the borrowers have paid back in full, so we are now looking at the second half of the borrowers, the older ones, and we have 48p in the pound for that. You get 100% from the first half and nearly 50% from the second half. That is why, overall, you get the sense that it is not that different from what you would expect from the overall cohort.

Q55            Lee Rowley: So, continuing in the spirit of simple explanations, Mr Roxburgh, can you give me a simple explanation of how I, as a member of the Public Accounts Committee, am supposed to conclude that this is value for money, given that, first, I don’t know who has bought them; secondly, I don’t know your value thresholds that said you could sell them; thirdly, I don’t know the model, because it was a black box, and fourthly, I have just been told that assessments of, for example, transparency elements have not been done? How can I confirm on that basis that these are value for money?

Charles Roxburgh: I will see if I can remember all four; I may not get them in the right order. Mr Manson addressed the transparency point. We do not think it is a material outcome, but we know that even allowing for that, we got more than these loans are worth to us. On the model point, it has been thoroughly audited by experts. The NAO looked at it, and their conclusion was that it was “reasonably accurate to date and within the…tolerance”, so it is not perfect but it is—

Q56            Lee Rowley: Based on one number—one set of data.

Charles Roxburgh: Yes, but that is all we have at the moment. The model has been tested and verified as best it can be. Your third point was that you don’t know what our actual retention value was. There is a good reason for that; we do not want to publicise it, because if the investors knew what it was, then they would just bid the price down. The NAO knows it, and we could share that in confidence if that would help to reassure you, but that is one instance where it is against the public interest to be transparent about our reserve price, because that would mean less money in the future. I’m sorry, what was your fourth point?

Q57            Lee Rowley: I don’t know who bought them.

Charles Roxburgh: That is the transparency point, but as Justin said, the overall majority would be blue-chip investors. We can share more detail in confidence.

Q58            Lee Rowley: Going back to one of Mrs Morris’s points, why have we not done a PSNFL assessment on this?

Charles Roxburgh: PSNFL is an important additional metric for thinking about the public finances, but in this case there is a particular problem with PSNFL, which is that it values the student loans at 100p in the pound. You can debate whether they are worth precisely what they are, and give different values, but everyone will agree that they are not worth 100p in the pound. They are not designed to be worth 100p in the pound, because they are designed for the people who do not go into the workforce and do not earn the threshold not to repay them, but PSNFL treats them as the face value. If you look at that and take off the face value of the asset, which is overstated, and the reduction in the debt, you get a very misleading impact from the PSNFL number. That is why, although we look at it in all our asset sales, in this particular case it hugely overstates the asset value, so it is not a helpful guide.

Q59            Lee Rowley: But that is no more misleading than the PSND, which usually understates the value of this. That is, if you are booking £1.7 billion into your accounts on a single year, that makes it look very good when actually, if your models are wrong, you could be losing an awful lot of public money as a result of doing that.

Charles Roxburgh: The factor to look at is whether the sale proceeds exceed the retention value, because that is the test of whether we are getting more money than if we held on to them and kept receiving the future payments. Because the sale value exceeded the retention value, it is value for money. The sale proceeds reduce PNSD one for one.

Q60            Lee Rowley: You have just explained to me the model of how that works. I get it and I accept it. You have explained it already. But the treatment here in PSND—a one-off payment of £1.7 billion into the accounts to reduce PSND and the deficit—is no more misleading than the PSFNL point that you were just making.

Charles Roxburgh: No, I disagree with that. PSNFL is looking at both sides of the Government’s balance sheet, and in the case of student loans it values them at 100p in the pound, which overstates it. If you look at the impact of this sale on PSNFL, you get quite a misleading effect, because it says that we have lost an asset worth 100p, which we haven’t, and we have reduced debt by the 48p that we sold it for, which we have. If you just look at the debt reduction, we got the 48p in the pound—the £1.7 billion—and we paid down debt by £1.7 billion. It is an accurate reflection of the debt reduction from the sale.

Q61            Lee Rowley: All on the assumption that 48p is correct.

Charles Roxburgh: Which comes back to the earlier point of whether that is the right value, and is it more than the potential value—

Q62            Chair: And to the 73p that Mr Slater’s Department had valued. On the PSND, we talked earlier about the reasons for sale and, obviously, paying down the debt. You talked about the significant impact over time. As more is sold off, you will be bringing in around £12 billion, so student loans as a whole on the Government books will account for a higher percentage of GDP. Was that a consideration when you were looking at loans, or was it all about paying down the debt?

Charles Roxburgh: At this stage, the only policy decision that has been taken is to sell plan 1 loans if it is value for money to do so. Even that is a material number, and the Chancellor has announced a target of selling £12 billion’s worth over the course of this forecast period. That is a material impact on the public finances if it can be done in a way that is value for money.

Q63            Chair: That is a target. Is it a target for the policy that we have discussed at the beginning of this session—to sell it off if there was no policy reason not to do so, but really, from the Treasury’s point of view, to pay down debt—or is there a consideration in Government thinking that as loans increase over time they will be a big percentage of GDP? Is that something that the Treasury or Ministers are thinking about?

Charles Roxburgh: If you look forward on the public finances, for a whole set of reasons, primarily demographic reasons—the OBR has set this out—absent policy changes, we face increasing public debt. The Chancellor has set out that it is really important that we continue to have debt falling. It is currently at around 85% of GDP. That is high, and it means that we would be vulnerable if there were a recession, and so the Chancellor has set out a fiscal strategy to bring that down steadily. That is the fiscal strategy. That has been our—

Q64            Chair: So it is about debt. Just to be absolutely clear, this and future sales are primarily about paying down the debt.

Charles Roxburgh: It’s about meeting our fiscal strategy. It is not that we use this piece of money for paying down debt; it goes into the overall assessment of the very significant increase in investments that the Government are making, with a big increase in capital investment. That has to be paid for. We also need to bring debt down and invest more, and asset sales have a role in both higher investment and reducing debt.

Q65            Chair: One of the other uncertainties in place is that the Office for National Statistics is deciding whether or not to overhaul the public reporting of the cost of student finance. Mr Roxburgh or Mr Slater, how are you factoring this in, and what are the risks to the proposals of Government to consider selling off future loan books?

Jonathan Slater: Either of us can answer it, but go on, Charles.

Charles Roxburgh: First, the ONS is independent, so it will do its review and we will apply what rules it comes up with.

Q66            Chair: Yes, but what are the risks to the Government’s plans?

Charles Roxburgh: For the sale proceeds, the value of these loans is actually independent of their treatment in the national accounts and their treatment on the departmental balance sheet. The value of these loans depends on whether and when the cash flows are paid back by the graduates and our discount rate, which is set by the Green Book.

The actual retention value of these loans is independent of the accounting treatment or the classification treatment, so the economic case for selling is unchanged by classification changes. That said, in the national accounts obviously we will apply whatever new rules the ONS applies, just as when the international accounting standards change we apply those.

Q67            Chair: So basically you are saying, in simple terms, for sales of loan books it makes no difference however they are classified.

Charles Roxburgh: The cash-flow value is unchanged.

Q68            Sir Geoffrey Clifton-Brown: What level of student loan subsidy is acceptable, Mr Slater?

Jonathan Slater: Student loan subsidy?

Sir Geoffrey Clifton-Brown: Yes.

Jonathan Slater: Whatever Parliament wants it to be. Whatever the Government want it to be. It is possible for taxpayers to pay the full cost of students going to university, as they used to do, which would be a 100% subsidy. It is possible to transfer all that cost to graduates—a 0% subsidy. There isn’t a right answer. It is a policy choice, isn’t it? Who do you want to pay the cost for people to go to university?

Q69            Sir Geoffrey Clifton-Brown: That leads me to the next question. The Chair has already referred to the ONS review. The Prime Minister has ordered a review. Some of these are to do with technical accounting or reporting, but some are presumably also to do with things like what the rate of interest should be on student loans. That will affect the timing on the future sales of these loans, won’t it, because you would not want to sell them if, in the next breath, you were suddenly planning to increase the interest rates?

Jonathan Slater: The Prime Minister’s review is a review of plan 2 loans—the £9,000 loans. We are handling a plan 1 sale programme of £1,000 to £3,000 loans. The Government has not even decided if it wants to sell plan 2. Any such decision will certainly post-date whatever comes out of the Prime Minister’s review of plan 2 loans.

Q70            Sir Geoffrey Clifton-Brown: What about the level of interest rates? If interest rates go up, at what level does that make the sale of student loans unviable?

Justin Manson: I think the NAO Report refers to several percentage interest rate points—I have forgotten the exact phraseology—and that is exactly correct. The reason we have not been more specific is, again, if we were to do that, investors in the next round would know what the Government retention value was. The judgment was that it was better that we did not share that in the public domain with investors.

If we look forward, the rise in interest rates to close that gap and close the VFM gap, if I can put it like that, is without any expectations from the marketplace at the moment. I think the market is looking at, over the next five years, the 12-year gilt rate moving something in the order of half a point, and it would need to be several points before the VFM gap closed.

Q71            Anne Marie Morris: Mr Slater, what will the impact be on your Department? You are managing assets that the Government does not own.

Jonathan Slater: The practical impact is on the Student Loans Company and other arm’s length bodies. There is an administrative impact on them in the part of the process of making payments to the investors. So we had to factor in the administrative burden on them of being able to make the process work. We did so, and the first set of payments went smoothly, so I think we have got that factored in.

Q72            Anne Marie Morris: Effectively, do you have separate “Chinese walls” between the administration of those assets you own and those you do not? While all is wonderful at the moment, if going forward something does happen and there is a change in Government policy, someone will have to negotiate with the current investors and you need to be clear which bit of your asset is which. Otherwise, you will not be able to deal with all of this.

Jonathan Slater: That’s one of the things that the Student Loans Company has to do—to be able to administer those loans separately from the others. There is a detailed compensation scheme that was built in, as you would expect, before those deals were signed.

Q73            Anne Marie Morris: How much cost will that add? If you keep selling off student loans—goodness knows what else you might sell off—how complicated will it become and how will you maintain all those Chinese walls?

Jonathan Slater: There’s not so much a Chinese wall. There is the need to calculate income streams due to investors. That is what we would not have been doing otherwise.

Q74            Anne Marie Morris: Isn’t there an issue about liability as well? Ultimately you are doing something for somebody else and not for yourself—i.e. the Government.

Jonathan Slater: Sorry, I am not sure I understand the question.

Anne Marie Morris: As I understand it, these loans have been sold off risk-free in the sense that you are continuing to manage the book and you are collecting in the sums and so on, but you do not own those assets; they are owned by somebody else. So there must be some contractual arrangement between you and them. There are also issues around the law: if something goes wrong, that will mean that you have to have managed and looked at things in a way that is not just about numbers and IT systems, but liability for doing a job for a third party.

Jonathan Slater: We have an arrangement in place, which Justin or Charles could talk you through if you would find that helpful, called a master servicing arrangement. It works out clearly that if Government were to fail to collect these loans—we have not transferred that responsibility—for a certain period of time, compensation would apply. We have built that into the regime and it needs to be administered. Arrangements will be put in place, overseen by a senior civil servant, to do so. So yes, there are administrative tasks in place that would not have been there otherwise.

Q75            Anne Marie Morris: Mr Manson, are you happy with that reply? Is it adequate given this is not just about numbers and administration, but about good governance and corporate risk—in this case, Government risk?

Justin Manson: I am very much satisfied with his answer. If I understand correctly, what your question is getting to is what the risks are of administration of these loans going wrong and what compensation would be due to investors if something did go wrong.

Anne Marie Morris: Yes, but also, unless you separate the loans, as opposed to them all being together on a computer and you then have to look at it after the fact, that might give rise to some legal complications. If you do it in advance and you have these Chinese walls, you may not run into them.

Q76            Chair: Do you need to have them in silos, because the Student Loans Company is administering them all?

Justin Manson: I don’t believe that we need to have them in different silos. I do not think it would be typical of other kinds of securitisations to separate them out into different silos and have them separately administered.

Q77            Anne Marie Morris: Even when they are differently owned? Surely, this must be a slightly odd situation where an organisation—in this case, the Government—is operating some assets that it owns and some that it does not, and is pushing them all together and changing a few buttons. It does not sound to me that the corporate governance is quite right.

Justin Manson: The incremental steps that the SLC has to do to administer these loans are relatively minimal.

Anne Marie Morris: That I accept, but the problem is when it goes wrong, because these are not the Government’s assets—they are someone else’s.

Q78            Chair: I think what Ms Morris is saying is that we all get casework about problems with the Student Loans Company—usually at the early end but sometimes later on. If there is a problem with the administration, and HMRC takes out the wrong amount—

Jonathan Slater: It is important for the Student Loans Company to know which loans we are talking about. That has been built into the system, and it is the 2002 to 2006 ones: they need to be identified separately, of course. There needs to be a contractual arrangement in place if something were to go wrong. Clearly, we would not have signed a contract with the sellers if there were to be a complaint to an MP about something going wrong—it is a much more significant event over four years, or if RPI has not been replaced by something else. In significant events, compensation is applied. We need to be able to define that, and that is what we did in signing the contracts.

Q79            Anne Marie Morris: Mr Manson, having looked at how those contracts are worded, is the way they are worded sufficiently robust to protect the Government, and therefore the taxpayer, should any claim be made against them?

Justin Manson: Yes they are.

Q80            Anne Marie Morris: What tests have you applied? Who did you ask for advice on that?

Justin Manson: The vital element is how well the SLC performs its job, and most importantly, how HMRC goes about collecting the money. In a typical securitisation, it is the collection of the money that is, in one sense, the most critical issue. Here, obviously we are very confident because we have HMRC doing the collections for around 96% of it. Then it comes down to questions of administration. We have spent our own time with the SLC, and the rating agencies and investors have met the SLC. All of them have been confident that the SLC is able to quite adequately manage the processes within the tolerances set out in the contracts that we have with them.

Q81            Anne Marie Morris: What sort of tolerances are in those contracts?

Justin Manson: Maybe I can give you the extreme example. If there is a problem reconciling cash flows, creating a servicing event that cannot be remedied within three years, that is a repurchase and compensation event. That provides a lot of latitude.

Q82            Lee Rowley: I am going to take you back to the model again for a moment. Somebody said a moment ago that there was not that much evidence of conflict within the process as it went through to realisation. It seems from afar that it was in everyone’s interest to make the models work. You had willing sellers, which you do not have to name. You guys wanted it off your books, particularly in the Treasury. What QA process did you go through to ensure that you were properly selling it for the right amount and that these models were the right ones?

Jonathan Slater: At its heart, what we are doing is an enormous amount of detailed analysis of the retention value. Alongside that, we were looking at what you would expect to get in the market. We did two different versions of each of those, because there is no exact number. Each of those numbers gives you a range. I am not going to be happy selling at the bottom of the range. That is work that involves the best experts within the civil service and lots of external advice, too. The NAO Report identifies essentially that it is possible, given the Green Book, to secure a win-win. The estimate from the NAO is that investors will get a return of 6.5% on average—obviously it varies. The Government get more than the retention value, because of the way that we can use the money we get, given the constraints on borrowing, which is the dispute with the Financial Times, and because we transfer the risk to the private sector. In a way, looked at in the way that the NAO has done, it should not be that surprising that it has been possible to sell, because we get a price that is higher than the retention value as defined by the Green Book, and they get a great return.

Q83            Lee Rowley: So it is all down ultimately to the happy path—the paradise scenario—that is based on your model. It is based on the assumptions that you have placed in it, so those assumptions have to be right. Out of that black box, can you give me some indicators that you used, such as long-term economic growth or population?

Jonathan Slater: Really, you should not be concerned that it is a black box, as you say. I spent a lot of time understanding it in detail, as did the Government Actuary, and we shared it all with the National Audit Office.

Q84            Lee Rowley: I am grateful that you did, but I will be concerned by it, so can you tell me what the economic indicators were?

Jonathan Slater: All sorts of indicators are used.

Q85            Lee Rowley: Long-term economic growth—what was the number?

Jonathan Slater: I see. What was the actual percentage used? Justin, do you know?

Justin Manson: I cannot tell you that off the top of my head, I’m afraid.

Q86            Lee Rowley: Give me something you could tell me off the top of your head. Give me some confidence that this black box on which absolutely everything hinges—

Justin Manson: For me to be able to be precise about this, I would need to have the line items you want to hear about, and I would need to see the numbers year by year.

Q87            Lee Rowley: Switch it around, then. If you do not have the specifics I am interested in, give me some out of your head that you think give a reasonable suggestion that this is a model that stacks up.

Justin Manson: Let me come back to the original point that you were making, which was, “Isn’t this all one-sided? Everybody wanted it to come out with the right result.” The rating agencies are there to test the optimistic assumptions, because it is not in their interests to have their credibility undermined by having an over-optimistic model out there that does not deliver. They test quite rigorously. In one case, they built their own model to test against the Government’s model. In the other case, they scrutinised the Government model alone. There is that significant check on it. Debt investors come with the same kind of perspective, which is that they will test the optimistic assumptions that we have. There is some check and balance in it, just to try to address—

Q88            Lee Rowley: But the fundamental problem with that is that we have to wait many years, and we only get a data point once a year to assess whether or not that is correct. Given there are very few people in the City and in Fitch and elsewhere who stay in the same job for 15 or 20 years, we have no institutional memory to know whether this is correct or not. The entire discussion we have had today is on the model. Of course that is the case. I cannot argue with that, otherwise we would do not any sale, but I cannot see how the QA processes—I can’t hook on to some confidence that you are trying to give me that demonstrates that it works.

I am asking you to give me some examples out of that box about why I would see, by the time I am 72, that the assumptions you have made are accurate, because I don’t know what I’ll be doing when I am 72; I have no idea how you guys would know what I’m going to do when I am 72. That is me, as a representation, as somebody who has actually had one of these loans—I was in this category. 

Justin Manson: To bring it to a simple data point, at the heart of the model that the Government uses here is taking an individual and saying, “What did you earn last year? What do we predict you’re going to earn next year?” And there is a set of data that we have, past data from previous student loans; we have that data. In addition to that, where we are missing data, particularly in forward years, we can use people who—I mean, the average age here for these borrowers is about 35. So how do we predict the 40-year-old, 50-year-old, 60-year-old type of transitions? For that, we rely on historical HMRC data, which is obviously the largest possible form of data we could look at to make those estimates. 

It is hard to disagree with the fundamental point you are making about the model’s accuracy—how do we guess what you are earning when you’re 72? We can’t do that. But the transition matrices, which lie at the heart of this—that is, what’s the probability that you will see your pay increase next year—those, I think, we can estimate with a reasonably high degree of accuracy across a population of borrowers.

Q89            Chair: But are you also tracking? You are looking at data from HMRC. But there is a range of difference in what a 25-year-old will earn next year, depending on what degree they did and what university they went to. Is any of that in this black box model?

Justin Manson: For the purposes of the model that we use for the sale, we looked at whether the university you went to and what course you did make a difference. They do, but by far and away the most important predictive factor for these cohorts, because they are relatively mature, is what they were earning last year. So we have stuck with that. If we were looking at a 24-year-old borrower, that might be a less good predictor, but by the time the average age has got to 35 it becomes quite accurate.

Q90            Chair: Interesting policy implications potentially for your Department, Mr Slater, but I think we won’t go down that avenue—

Jonathan Slater: That’s precisely why we need to have a different value for this, because they are the oldest ones, not the average.

Q91            Chair: I will come to Sir Geoffrey in a moment, but I just wanted to ask a question about looking ahead to the future. You might have seen some of the evidence that went to the Lords Committee. One of the issues is that where Government policy can change quite easily—where a Chancellor or Government can choose to increase the threshold for repayment—that goes down very well with students and other people who are subject to these loans. But according to some of the research—it was actually submitted by Professor Barr from the LSE—the impact of an increase in threshold from £21,000 to £25,000, which was the April 2018 increase, has no effect, of course, on debt, but increases the fraction of borrowers who will not repay in full from 77% to 83%, so that only 17% of borrowers will repay in full.

When you are looking at the future loans—this question is really for you, Mr Slater—how are you working through all the figures so that you can advise Ministers who might be making those sorts of policy announcements about their impact, because it is not very much of a problem today but this is a big issue about what’s coming back tomorrow, and if you’re then going to sell it off, presumably that has a big impact on the valuation of the loan book.

Jonathan Slater: If the Treasury was to come to me with a proposition around the selling of the plan 2 loans, or if they haven’t—I won’t predict the future—

Q92            Chair: We are working on the premise, as Mr Roxburgh said, that it is an intention that has been announced.

Charles Roxburgh: I said that no decision has been taken—

Q93            Chair: No decision, but it is a general direction of travel that the Government will explore—

Jonathan Slater: I don’t think the Government have expressed any view on the subject.

Chair: I think Ministers have—all right. Okay.

Jonathan Slater: Then obviously we would need to do a fantastically detailed analysis of the sort we had done for plan 1 loans, but for the plan 2 loans, and work out what a value for money proposition was, whether we thought we would get one, and so on and so on. That’s a separate question from the question we were exploring a little while ago about the balance between graduates and taxpayers. You are right that a significant decision was taken last year to ask taxpayers to make a bigger contribution towards the cost of students going to university, using that threshold increase as an opportunity. That is a political policy choice, isn’t it? The Government will have to make some policy choices on the back of the review by Philip Augar, which is going on at the moment. That seems to me to be a separate question in kind from the question of, if you have got some loans and you continue with the loan system, on any given set of loans, is it better value for money to sell them or keep them? That is a separate question.

Q94            Chair: But if you were going to sell off the loan book and you increased the threshold, you are actually reducing the value of the loan book to purchasers.

Jonathan Slater: You would generally want to avoid a situation in which you were selling something and then messing about with it straight afterwards.

Q95            Chair: Right. There are two questions. First, if you do it beforehand, you potentially reduce the income to the taxpayer because of the threshold increase. If you increase the threshold, you are reducing the value of the loan book because fewer people will pay back the full loan.

Jonathan Slater: Yes. As I say, that’s just another way of saying that you want the taxpayer to pay more for university. That’s one political choice.

Q96            Chair: What about after the sale?

Jonathan Slater: Separately, with a set of loans, there is a value for money debate about whether to keep them or not. There would obviously be a cost associated with compensating people who have bought loans if you were to change the terms and reduce their future income stream. Obviously, I would want to avoid a situation in which that happened, but equally I can’t fetter the discretion of future Governments to do what they want.

Chair: No, no. Absolutely not.

Q97            Sir Geoffrey Clifton-Brown: On the back of this, you have got the ONS investigation on financial reporting and the Prime Minister’s investigation on the level of tuition fees. When are they likely to conclude?

Jonathan Slater: As Charles said, the ONS is obviously an independent body. It is engaging with Eurostat—that’s going on at the moment. I’m not certain, but I would expect some response to be coming from the ONS in the next few months. Would you expect that?

Charles Roxburgh: Yes, but it’s a matter for them.

Jonathan Slater: I’m not in charge of this process, but I’m expecting the ONS to come back in the next few months. In respect of the post-18 review, the terms of reference that were published earlier suggest that the Government is going to complete it in early 2019. We can have a debate about what the meaning of the word “early” is if you would like. It is no more precise than that.

Q98            Sir Geoffrey Clifton-Brown: Okay. You say that a decision about whether or not to sell off the loan 2s hasn’t been made. Presumably a decision cannot be made for the very reasons you’ve set out just now—you don’t know exactly how you are going to shape student loans until those reviews are complete and Ministers have had a chance to consider them.

Jonathan Slater: We’ve got a lot of work to do to complete the plan 1 programme, which runs to 2022. We’ve got loads of work to do before we get to that stage.

Q99            Sir Geoffrey Clifton-Brown: Nevertheless, the Government’s direction of travel is to want to raise £12 billion from the sale of these student loans.

Jonathan Slater: The plan 1 loans. There is £43 billion-worth of face value, so there is enough in plan 1 to generate £12 billion.

Q100       Sir Geoffrey Clifton-Brown: So they can proceed independently of the loan 2s?

Jonathan Slater: Exactly.

Q101       Sir Geoffrey Clifton-Brown: And when will a decision be made on that?

Jonathan Slater: On what?

Q102       Sir Geoffrey Clifton-Brown: On sale of the remaining tranches of the loan 1s.

Jonathan Slater: As the NAO Report identifies, the £12 billion works on the basis of at least one sale a year for the next few years. Clearly, precisely when one sells and how much depends upon market conditions. We have got that sort of programme between now and 2021-22.

Q103       Sir Geoffrey Clifton-Brown: It looks as though the NAO were predicating the sale of about £3 billion a year for the next five years—in other words, £12 billion to £15 billion. Are those the sorts of numbers that you are considering?

Jonathan Slater: £1.7 billion is the first bit of the £12 billion, so we have got another £10 billion to go by 2022 if we hit the OBR’s estimate, but obviously what happens in practice will depend upon market conditions at the time and whether there is a value for money proposition on the basis of one or more sales a year between now and then.

Charles Roxburgh: £12 billion is not assuming selling the whole plan 1 book. There are still some more left unsold after that £12 billion. You could extend it and add to that £12 billion. That hasn’t been taken as a decision yet.

Q104       Chair: These are the last couple of questions on the transparency point. Mr Manson, on the companies that did not want to be named, can you give us specific reasons as to why they did not want to be named?

Justin Manson: I don’t think I would add anything to what I said before.

Q105       Chair: They felt they were commercially sensitive.

Justin Manson: Exactly.

Q106       Chair: Commercially sensitive about the actual purchase or about their reputation?

Justin Manson: I don’t think it had anything to do with the reputation.

Q107       Chair: Mr Roxburgh, we know the rough range of the retention value. Given that the purchasers have got the model and they know the principles and discount rate for calculating the retention value, surely the brains in those sectors are able to work out roughly what the retention value is, so why the secrecy of the retention value?

Charles Roxburgh: They don’t know our assets’ specific risk. They might be trying to work it out for themselves.

Q108       Chair: How close, within tolerance?

Jonathan Slater: You asked that question. Anything that we offer you of a quantifiable nature gives bidders more information about what value we achieved from it. We are happy to offer you information in commercial confidence. There is no lack of transparency with the Committee, but we cannot possibly give you any information in the public domain.

Q109       Chair: They will be getting pretty close, I guess.

Jonathan Slater: I can’t offer you a view on that because it would reduce the price next time round.

Charles Roxburgh: And it’s very important to maintain competitive tension in these bids. They may think they can work it out, but they will work it out differently. That helps to get more tension in it. We think it is really important that we keep the sale price as far north of the retention value as we can.

Q110       Chair: We as the Public Accounts Committee would agree with that. Sir Amyas, do you want to add anything?

Sir Amyas Morse: When you are building the book and so forth, if it is possible for these major investors to know the principles and all the criteria you are using, they will have a pretty good stab at calculating the price. Your answers to us imply that you think the fact that that remaining element of uncertainty or whatever it may be has the effect of driving up the price. You clearly believe that or you would not be secretive about it. There must be some benefit to not simply saying what the retention value is.

Justin Manson: Could I answer from my narrow perspective?

Sir Amyas Morse: I don’t quite get why we are giving the game away by asking if there is a principle of confidentiality. You are assuming that that principle yields something in terms of a better price.

Charles Roxburgh: The confidentiality adds to the retention price absolutely. If you say, “These loans are worth x to us,” people know that if they bid above x, and others might bid a bit more, they know the point at which we will sell the assets, and that gives away a lot of our negotiating position in an auction. Absolutely. Mr Manson was asking different questions about confidentiality for a subset of the investors and what the quantification of that was. There are reasons why, if you want more investors and more people bidding, but on the confidentiality of the retention price we absolutely think that we should keep that quiet to get the best price for the taxpayer.

Sir Amyas Morse: So if we look at this over a series of transactions over some years, we will be able to assess how well that strategy has worked by seeing whether on average you have managed to get far north of the retention price in these sales. Is that fair?

Justin Manson: I would add a word of caution, if I can interject. In a reasonably bullish market environment such as the one we had last year, it might be one indicator of success. I would only say one indicator. In a less bullish environment where the difference between the sale price and the retention value was very narrow, I am not sure that would be a good test of how successful we had been. On the contrary, it might be that in a very difficult market the Government had got a very good price for the taxpayer.

Q111       Chair: It is interesting. We might have a discussion with the National Audit Office. We have been discussing in preparation for this a lot of the asset sales. We might have a conversation about how we measure retention value. The bits that the NAO can see, we can see privately too. We will park that one for now. I want to bring in Sir Geoffrey for one last point.

Q112       Sir Geoffrey Clifton-Brown: Given your earlier replies, Mr Manson, and given the figure on page 15 that there are 59 investors, if you add those numbers up, and you are having post-investor discussions to see how they have evaluated the bids, can we take it that one of the items in those post-investor discussions will be the issue of confidentiality? Surely you will want to do that to assess, in your words to me in response to my earlier question, that we will impose exactly the same condition in the next tranche of sales. Obviously, in your mind at the moment, you are thinking about keeping this confidentiality, but you have heard the concern of the Committee today, and you will want to review that position quite carefully. One of the aspects that you will want to review is the investor sensitivity to the matter.

Justin Manson: Yes, it will be.

Q113       Chair: Great; I love quick answers. One other point: £12 billion is projected for the remaining tranche of loans. How robust is that figure?

Charles Roxburgh: You say how robust. That was the forecast that the OBR put in their last projection.

Chair: So it is an OBR figure. Does it match with yours?

Charles Roxburgh: That is based on our sense as to what is a realistic programme. We have talked to UKGI. Obviously, like any projection of asset sales, it depends on market conditions at the time. Likewise, they project forward an estimate for selling RBS shares, but again, that is a—

Q114       Chair: Have you modelled it? If Brexit goes badly, and the market is less buoyant than the one Mr Manson described, have you modelled for that and have you modelled for a bounce in the market if everything goes brilliantly?

Charles Roxburgh: Well, the Government’s policy is that we will have a successful Brexit. The OBR forecast is based on the assumption of a smooth Brexit. Therefore, the asset sales are—

Q115       Chair: We did have a referendum where there were not even six civil servants in the basement of your Department, Mr Roxburgh, working out what would happen if the vote went to leave. Following that lesson, are there not even half a dozen civil servants working out what would happen to the £12 billion figure—among other figures, but that is the one we are talking about today—if it went badly? There must be some modelling that you are doing. Government is always pessimistic from my experience.

Charles Roxburgh: Across Government, there is extensive work on how to prepare for a no deal—that has been made very clear. There is extensive work—

Q116       Chair: Does that include modelling? How robust is this figure? What is the tolerance level of it?

Charles Roxburgh: Clearly, an outcome that nobody wishes to see—a disruptive no deal—could have implications for financial markets. It is not for me to speculate on what those are. Market conditions and appetite for assets would then be different, and one would have to revisit those forecasts. But we are not doing that. We are working on the basis, for these projections, of the OBR forecasts. Separately, yes, as you would expect, contingency planning is happening across Government for no deal, as is public.

Q117       Chair: My concern is that these figures get banked before they are actually delivered, and there is a lot of—

Jonathan Slater: No, no. The process, as we said from the very beginning, is that we have to identify whether or not it is value for money. The fact that the OBR thinks the £12 billion is going to be delivered by 2022 is their best forecast—they have been set up for that process to be independent. The amount that will actually be banked will depend on value for money, which will depend on how much bidders bid for it. In reality, whatever they judge, on the basis of the facts at the time, will be what we advise Ministers.

Q118       Chair: And also, it will depend whether you sell it.

Jonathan Slater: Yes. If we do not sell it, we do not sell it, and if we do, we do.

Q119       Chair: So you would consider holding off if you felt you could not—

Jonathan Slater: It would be for the Government to decide, but obviously, if it were not value for money to sell, I would recommend not to sell it.

Chair: I say you—I mean you as a representative of Government.

Jonathan Slater: I as the accounting officer would recommend against selling something if it was not value for money—not in accordance with the way the Financial Times does it, but with the way the Treasury does it.

Chair: We will leave it there. Thank you very much for your time. The transcript will be up on the website, hopefully by Wednesday, thanks to our good friends at Hansard, and we will be producing our report around October time, because we will not be here now after this week. Thank you very much indeed for your time, and we look forward to getting a response to our report.