Student loans are in principle a straightforward business. The government lends students money; after they graduate, they begin repaying it. From the perspective of politicians and the Treasury the advantage of loans over grants is clear: the money isn’t simply given away, it comes back over the lifetime of the loan. Even better, in the national accounts the loans are classified as ‘financial transactions’, not ‘expenditure’, and are excluded from calculations of the deficit. When in 2012 the coalition all but ended the direct-grant funding of undergraduate teaching in English universities and colleges, the move could be sold as consistent with fiscal austerity – it had the effect of reducing government spending. But the income of universities and colleges was spared the cuts made elsewhere because the gap was more than filled by higher tuition fees backed by loans.
Since 2012 English higher education institutions have been able to charge new full-time students from the UK and EU up to a maximum of £9000 per year for tuition. Anyone graduating from 2015 onwards is likely to owe £27,000 in tuition fee loans and more for maintenance loans, plus whatever interest accrued on the loans while they were studying. The Institute for Fiscal Studies reckons that on average student borrowers will owe £44,035 at graduation; for those who began their degrees before 2012, the figure was under £25,000.
As the general election approaches, both Labour and the Conservatives are committed to further austerity. George Osborne has made clear that under a Tory government ‘fiscal consolidation’ will be maintained until 2019 at least. Where further savings and cuts – roughly £55 billion over the next parliament, ‘cuts on a colossal scale’ according to the Institute for Fiscal Studies – are to be found hasn’t been made clear, still less where the money will come from to pay for the £7 billion in tax cuts David Cameron airily promised in his speech to the Conservative Party Conference last October. When such pledges are discussed in the media, it is usually in terms of whether they are funded or unfunded – whether the numbers ‘add up’. Osborne’s Autumn Statement in December surprised analysts and angered the business secretary, Vince Cable, who demanded that the independent Office for Budget Responsibility set out Tory and Liberal Democrat spending plans separately. The OBR reported that planned cuts to departmental spending outside health, schools and overseas aid ‘would pose a significant challenge if they were confirmed as firm policy’.
But if the government – any government – is to achieve the desired reduction in the deficit, it will need to look beyond tax and spend. The most obvious means of raising cash by other means is a sale of assets. Here, the clear candidate is the student loan book. A million students take out loans every year. When the coalition was formed in 2010 the total student loan debt owed by English students and graduates was about £30 billion. That debt reached £54.4 billion in March 2014 and is set to increase rapidly over the next few years: the cash outlay on loans is now £10 billion a year and is projected to have passed £15 billion by 2019-20, by which time annual repayments will barely touch £2 billion. Because of the ‘generous’ conditions on the loans, repayments won’t reach significant levels until the 2030s. If things continue as planned, the sum that graduates owe will peak in the 2040s at somewhere around £330 billion in today’s money: a sizeable asset. However, the Treasury is acutely aware that so long as there is a shortfall between annual outlay and repayments, the government must borrow to create new loans in the first place, and that adds to the national debt. The particular lure of selling student loan accounts is that it would allow the government to swap repayments scheduled to come in over the next 35 years for cash today. Cash is always useful, and long-term repayments do not help governments meet the five-year deficit and debt reduction targets on which their economic competence is judged.
The real mystery is why the sale of the loan book hasn’t been set in train already. The last Labour government put the legal framework in place, and all three mainstream parties entered the last election pledging to launch a sale. Once in government, the coalition appointed the financial advisers Rothschild to run a feasibility study. In 2011, the higher education White Paper, Students at the Heart of the System, was bullish about the options. Then there was a hiatus, broken only in December 2013 when Osborne announced that a five-year auction process estimated to raise £10-12 billion in total would begin in 2015. That sum was much lower than originally anticipated and involved the disposal of loans issued more than ten years ago, not those issued from 2012 onwards. Just seven months later, in July last year, Cable announced that he and Nick Clegg had agreed not to proceed with the sale. (As secretary of state with ultimate responsibility for universities, the decision whether or not to sell rests with Cable.) Not that Osborne and the Treasury have been dissuaded. Osborne announced in the Autumn Statement that ‘progress continues’ on the planned sale, and the projected £12 billion gross proceeds are still included in the OBR’s fiscal projections.
This will be a relief to many, since Osborne promised in 2013 to use the cash from the sale to expand the number of funded undergraduate places by 30,000 in 2014-15, as an interim step towards removing the cap on admissions altogether the following year. As it is, only 15,000 additional places were taken up this academic year and the government’s estimate of how many places would be filled once the cap was lifted has been quietly revised down from 60,000 to 45,000. These revisions have allowed Osborne to expand loans to fund postgraduate study for the under-thirties without adding to the planned cash outlay.
So what is going on? If the sales have been central to government thinking about student loans, and continuing austerity makes the disposal of assets a more attractive idea, why the hold-ups and hiccups? Why, when Cable vetoed a sale this parliament, does the Treasury still think it can book some cash by March 2016? It’s quite difficult to get to the bottom of things. Most of my requests for information have been rebuffed with appeals to commercial confidentiality: disclosure might ‘adversely affect’ the government’s negotiations with potential purchasers. But I do have a copy of Rothschild’s study, Project Hero: Updated Views on Feasibility, which gives us a key to interpret the coalition’s to-ing and fro-ing.
Student loans were first introduced in the UK in 1990 to replace maintenance grants. Borrowers who signed up to these ‘old-style’ or ‘mortgage-style’ loans repaid them in sixty monthly instalments once their income crossed a threshold of 85 per cent of mean annual earnings. The size of the repayments was determined by the total amount borrowed. It was felt that these terms were unsuitable in light of the expansion of the sector in the 1990s and the larger borrowing that would be needed for tuition fees.
In 1998, the UK government introduced a new kind of loan. The income threshold for repayment of the Income Contingent Repayment (ICR) loan determines not just when repayments start but also the amount repaid: 9 per cent on all earnings above the threshold (£21,000 from 2016 for the most recent loans). That is, repayments are determined by income, not by the amount owed. One significant feature of ICR loans is that the mandatory monthly repayments will often be lower than the interest accruing on the outstanding balance. As with other loans, the repayments end if the balance is cleared; but the government also commits to write off any balance outstanding thirty years after the loans become eligible for repayment. There is likely to be an outstanding balance in the majority of cases. Talk of ‘default’ mistakes the nature of the loans: graduates may eventually have their balances cleared by the government, but that doesn’t mean they have defaulted on the terms of repayment. The initial estimate was that fewer than 40 per cent of borrowers were expected to clear their accounts. The IFS now estimates that roughly three-quarters of borrowers will have at least some of their debt wiped out. The government expects to lose money on each cohort and in this sense continues to subsidise higher education.
In this respect, ICR loans are very different from commercial loans. Politicians encouraged students to think of them as being ‘like a tax’. In its efforts to sell them on, the government faces the difficulty of pitching this ‘novel asset class’ to potential buyers. What’s more, it is difficult to value unsecured loans – no collateral is needed to get a student loan – with such long lifetimes and unfamiliar conditions of repayment.
The ‘mortgage style’ loans have already been sold to the private sector. In the late 1990s, the Labour government sold two tranches for £1 billion each. The price matched the official valuation of the loans, but the government also agreed to compensate purchasers through additional annual payments if the number of borrowers crossing the earnings threshold was lower than expected or if there were adverse movements in inflation and interest rates. The contracts will run through to 2028 and government losses on the deal so far amount to £240 million (about £160 million in 1999 terms). That is, the government is today £240 million worse off than if it had kept the loans in national ownership. It should be emphasised that this was anticipated: the government of the day entered into the sales expecting to have lost £140 million by the time the contracts closed thirty years down the line. This is what Martin Wolf had in mind when he wrote in the Financial Times in 2011 that selling off the loan book was ‘surely economic illiteracy. The loan book is sure to be most valuable to the state, which has much the lowest cost of funding’ (the government has much lower costs of borrowing than a private fund or company as it is thought less likely to default).
The government would always have lost money on the sale in the long run, whether through price or compensatory payments, but receiving £2 billion up front helped Labour operate within the fiscal strictures set down by Gordon Brown in his ‘golden rules’: spending must be balanced by taxes over the economic cycle and the government should borrow only to invest. This immediate benefit of the loan sale was seen to outweigh future losses.
Recently the coalition government sold the last of the mortgage-style loans for £160 million and claims to have made £12 million on the deal. But the earlier sales set a precedent: the money generated today by asset sales can justify a long-run loss on the deal. ICR loans are more complicated and the sums involved in selling them will be much, much larger, but the considerations are the same: value for money and long-run economic considerations versus the short-term pressure to generate cash without additional borrowing.
In 2006, student loans were extended to cover tuition fees, which universities and colleges were allowed at the same time to increase to £3000 per year. Labour announced its intention to begin selling off the loans in the 2007 Budget, and prepared the way with the 2008 Sale of Student Loans Act. This legislation enabled the secretary of state to make a sale without consulting the borrowers, and to ‘include provisions in any sale to compensate a loan purchaser in specified circumstances’. The plan was to sell off £6 billion of loans every two years. But the financial crash intervened and the sales were put on hold. Instead, just before the 2010 election, Labour put out a tender for a feasibility study on ‘alternative routes to market’. This was meant to run alongside the review of undergraduate funding chaired by Lord Browne, whose report is seen as having ushered in the coalition’s funding revolution. Although Browne considered recommending a graduate tax, his review came down in favour of extending the loan scheme. As a statement from the Russell Group, ‘Objections to a Graduate Tax’, pointed out at the time, it would not be possible to sell ‘graduate obligations’ unless they were structured as loans.
After the election, the coalition awarded the contract for the feasibility study to Rothschild. The higher education White Paper published in summer 2011 stated that the ‘full range of options’ would be considered, ‘including retaining the loans on the government’s books, selling them outright to financial investors, or selling loans to one or more regulated companies set up to manage the loans.’ Rothschild was given a clear remit by the Department for Business, Innovation and Skills to investigate a ‘repeatable’ model for an annual sale of new loans (since BIS is responsible for universities, student loans are recorded in its accounts). Borrowers should not be put in a worse position as a result of the sale (though whether or not they had been would be judged by the government), and the taxpayer should get ‘value for money’, assessed as a comparison of the ‘expected value of achieving the sale with the expected value of retaining the assets over their life’. The main aims of a sale, from the government’s perspective, were to reduce its exposure to the risk of non-repayment, and to lower national debt.
National debt has risen to £1464 billion (80 per cent of GDP), which is a long way north of the old ‘sustainable’ target of 40 per cent (Osborne described it in his conference speech this year as ‘dangerously high’). Part of his declared ‘fiscal mandate’ was to have national debt falling as a percentage of GDP by this year. (The OBR’s current prediction is that it will continue to rise until 2016, and will then begin to fall slightly.) It is now thought that the new higher education funding system will add more than £100 billion to the national debt before repayments reach a significant level in the mid-2030s. At that point, the OBR thinks, the borrowing to create student loans could constitute one-fifth of national debt.
These fantastical projections are for the future, but student loans have begun to impinge on current fiscal decisions. The IFS dismissed Osborne’s claim that you can finance new loans by selling old ones as ‘economic nonsense’. You cannot sell an asset for what it is worth to strengthen public finances. You can only change the timings: cash today or cash spread over thirty years and more. But such a sale can improve the headline debt statistics: you can make the nation’s financial health look better to the public in the short term. And cash from a loan sale can be set against the national debt or used for other spending.
It is from this perspective that we should consider the brief justification Vince Cable offered for his decision: ‘The government was considering the sale of student loans on the basis that it would reduce government debt. Recent evidence suggests this will no longer be the case.’ This evidence hasn’t been made public and does not appear to have swayed the Treasury. It is entirely possible that Cable has managed to delay plans for a sale only until after the general election, when a new secretary of state may well take a different stance. Conservative ministers still want a sale and Liam Byrne, Labour’s shadow minister for universities, indicated at last year’s party conference that he had been ‘pressing’ Cable to crack on, even while admitting that it might turn out Cable had made the right call. When Labour last took a line in public, in 2013, its position was that it opposed a sale only if it led to worse conditions for borrowers. The difference between the parties may come down to the degree of financial engineering involved and hence the amount of money that could be generated.
The Rothschild report, Project Hero, drafted in late 2011, reveals the broad outlines of what is at stake and how a sale might be achieved. It sets out the advantages and disadvantages of a number of proposals and explains that the option it initially preferred was abandoned after feedback from the Office for National Statistics. Certain schemes, it seems, would not be classified as sales because they do not allow a ‘full risk transfer from the public to the private sector’: either the liabilities associated with the loans would still be on the public books or the government would be entering into an ongoing commitment after the sale to make the deal more attractive to the private sector. (It is my understanding that the sales from 1998 and 1999 would no longer be classed as ‘sales’ by the ONS.)
Rothschild came up with an alternative proposal, which is in most ways consistent with what was announced by Osborne in 2013. Rather than selling each year’s new loan outlay as originally planned, the government would instead carry out a ‘retrospective’ sale of loans already issued. Purchasers would buy the debt of whole annual cohorts, starting with the first to take out ICR loans: those who graduated in 2001, 2002 and 2003. The government would continue to administer loan accounts and collect repayments on behalf of the purchasers.
There was a large hurdle. The likely bidders – pension funds and insurance companies – were wary of the interest-rate terms on these ‘pre-2012’ loans. Outstanding balances on these loans accrue interest at whichever is lower: the Retail Price Index (a standard measure of inflation), or bank base rates plus 1 per cent. Base rates have been 0.5 per cent since 2009, while RPI has been higher (around 2.5 per cent in 2014), so borrowers currently benefit from an interest rate of just 1.5 per cent. (The 2011 Education Act removed this clause for loans issued to students starting after 2011. They now face interest rates that range from RPI to RPI+3 over the lifetimes of the loans; there is no reference to bank base rates.) Rothschild told the government that only by circumventing this clause could a sale of between £10 billion and £20 billion be achieved; otherwise, it would be lucky to make £2 billion. Investors do not want assets that return less than inflation. ‘The risk is best taken by government,’ Rothschild said. ‘They can afford it, assess it and ameliorate it better’ than any potential investor. Second best, the report suggested, was that the risk be taken on by graduates themselves: ‘Their earnings are more likely to reflect inflation than base rate.’
The government had already ruled out the ‘second best’ option, which would require it retrospectively to raise the rates at which interest accrues against outstanding balances – with the result that graduates would take longer to clear their debt. This meant that the government would have to take on the risk itself if it wanted significant sales. It was advised to do this through a financial device known as a ‘synthetic hedge’. The purchaser would receive cash flows equivalent to those resulting from changed interest-rate terms for borrowers, but instead of graduates making the additional payments, the government would make up the difference. This gives the purchaser a ‘hedge’ against adverse inflation movements; the deal is ‘synthetic’ because, for the purchaser, it is as if the interest terms had been changed. In January 2014, David Willetts, then minister for universities, confirmed that the synthetic hedge was what the government had in mind, when he discussed this ‘investor protection instrument’ in public for the first time.
The government would ‘compensate investors’, Rothschild summarised, ‘by paying them cash flows (towards the back end of the maturity profile) representing the final payments borrowers would have made had the cap been removed.’ Final payments, notice, and therefore to be made by future governments, not the one entering the deal. It may be, though, that the synthetic hedge, seen as a new liability for government, was enough to make the ONS conclude that it wouldn’t allow for ‘full risk transfer’.
The whole business boils down to value for money, on both sides. The private sector may take on more risk but will want a lower price in return. What represents a fair price for student loans? ICR loans are unusual: there is nothing to compare them to in today’s debt markets. It isn’t so much that the loans aren’t run on a commercial basis, but that the majority of repayments are expected to materialise a long way into the future, which introduces more uncertainty into the cash-flow projections. With normal loans or mortgages, I know what my repayment commitments are (they are fixed in the initial agreement), even if, for example, I don’t know the extent to which inflation will erode the value of the repayments. With ICR loans, the repayments themselves vary according to factors such as inflation and the state of the graduate labour market, adding further uncertainty.
There is no market price that can be used to evaluate loans of this kind. Complex modelling is required to estimate the value of repayments thirty years and more into the future. By these means the government generates a ‘fair value’ – what it thinks the future cash flows generated by the loan are worth – which it records in the accounts. ‘Fair value’ differs from ‘face value’, which is the sum of all outstanding balances, or what is nominally owed to the government. The latest accounts from BIS, published last July, value its stock of loans at £33 billion based on a face value of £54 billion. That is, although student and graduate borrowers owe £54 billion to the government, it expects to receive only the equivalent of £33 billion in repayments.
The auditor general has repeatedly stressed the ‘number and volatility of assumptions underpinning valuation’. In February 2014, the Public Accounts Committee addressed the crux of the matter:
We were told that the value for money of a sale would depend on a comparison of how much purchasers are willing to pay against an estimate of what the loans are worth. But to make this comparison, [BIS] needs a reliable and accurate forecasting model so that it can make a sufficiently robust estimate of the loan value in the first place, which it has not yet been able to do with any confidence.
The National Audit Office had already told the government, in late 2013, that its modelling wasn’t up to scratch. A new model was implemented in spring 2014. For new loans issued in 2014-15, BIS now expects to lose 45p on every pound loaned, compared to an estimate of 30p when the policy was approved in 2010. When you are issuing £10 billion of loans every year, each penny in the pound is the equivalent of £100 million. As a consequence, the fair value of loans already issued was written down by roughly £3 billion. Such dramatic changes in valuation are almost certainly another reason the government is looking to sell. It is holding an asset of uncertain worth: perhaps it’s better to cash in now, even if the sale turns out to represent a greater loss over the long run? The difficulty is that the private sector is likely to put an even lower value on the loans.
How are such valuations made? Take a basic example. How much would you lend or pay for the promise of £100 in ten years’ time? In making this assessment you might consider estimates of inflation; of your own costs of borrowing if you can’t take up the offer with the cash you have to hand; of the returns on alternative options; of the risk that the £100 will fail to materialise in ten years’ time etc. Taking all these things into account, you come up with a ‘discount rate’, a measure that reflects what an IOU for £100 in ten years’ time might be worth to you today. It is normally expressed as an annual rate.
At a discount rate of 5 per cent, I would value £1 today as the equivalent of £1.05 to be received next year, £1.10 in two years, £1.16 three years from now, and so on. Conversely, £1 received next year would be worth 95p today; £1 received in two years, 91p; £1 received in four years, only 82p. On the basis of a 5 per cent discount rate, £100 received in ten years’ time would be valued at £61.39 today. The higher the discount rate, the higher the preference for cash today over future money. Rothschild indicated that potential purchasers would price student loans against alternative investment opportunities offering an annual return at least 3 percentage points above inflation (RPI). With RPI currently at 1.1 per cent, that would indicate a discount rate of 4.1 per cent, in which case the promised £100 in our example would be valued at just £66.91 today.
In calculating a fair value for its student loans, the government uses a discount rate of RPI+2.2; with RPI at 1.1 per cent, that’s 3.3 per cent. At that rate, the government values £100 in ten years at £72.28 today. This official figure was established in 2006, and has been criticised on the basis that it hasn’t been updated to reflect the significantly lower cost of government borrowing since the financial crisis. The rates on ten-year government bonds (‘gilts’) are currently 2 per cent. If that were the basis for the discount rate, then our £100 would leap in value to £82.03. That’s how important the discount rate is. The National Audit Office describes the BIS approach to discounting as ‘prudent’, but if you are planning to sell, then ‘prudent’ can easily come to mean ‘undervalued’; it’s possible that the estimated loss of 45p in the pound will prove to be overstated.
It is worth stressing one final striking point. In assessing the ‘value for money’ case for a sale, the Treasury has instructed BIS to use a different discount rate. Although the fair value of the loans recorded in BIS’s accounts is based on the discount rate of RPI+2.2, the sale would instead be judged against a discount rate of RPI+2.5, plus a further increment to reflect any premium the market may add for risk. This would bring the valuation of any future cash-stream in line with the private sector’s assessment of RPI+3. The Treasury insists that this instruction is consistent with policy: the loan sale would be a ‘negative spending decision’ and the higher rate better reflects society’s preference for cash now. One might feel that depends on what the cash is used for (Cameron’s tax cuts?). The lifetime of a student loan is long, and the bulk of the repayments will be made towards the end of it. The level of the discount rate therefore has a big effect. When billions of loans are at stake, the effect is magnified accordingly. For the £10 billion of new loans issued this year, a discount rate of RPI+3 as opposed to RPI+2.2 lowers their value by £660 million (according to the ‘ready reckoner’ on the BIS website). This amounts to a further loss of 6.6p in the pound on top of the 45p BIS already expects to lose on student loans issued this year. Given the controversy over the sale of Royal Mail (which was a one-off, not an annual event), the potential for large, repeated losses on annual loan sales may have had some influence on Cable’s decision. Again, after the election a different politician may be persuaded to take the Treasury’s view on the overriding need to shrink the balance sheet.
In July last year, just after he had stepped down from the cabinet, David Willetts went public with a scheme that would allow universities to buy their own graduates’ debt. (BIS released a terse comment: Willetts’s suggestion ‘does not represent government policy’.) ‘As more undergraduates at English universities take out student loans,’ he wrote in the Financial Times, ‘the government is rightly looking at selling this increasingly significant asset.’
If the asset is so significant, shouldn’t we consider holding on to it? The value of student loans as an asset is linked to graduate earnings: if you believe in the growth to come, why would you sell? The thinking here is very much like the thinking behind PFI: it is better to keep investment off the balance sheet even at the cost of a long-term loss. In his Autumn Statement of 2013, Osborne drew attention only to the short-term benefits of the sale. Yet it isn’t at all clear that a sale represents good long-term economic sense for the government or for a public that hasn’t been consulted, and which might be surprised to learn that a sale can proceed without a vote in Parliament.
 Although the annual cash outlay on new student loans is excluded from the current measure of the deficit, the annual interest paid on the money borrowed by government to fund the loans does count. Three decades from now, when the final losses on each year’s loan issue is known, those losses will count in national expenditure, but as a ‘capital transfer’. At present, capital transfers don’t affect the deficit, as it is based only on current expenditure.
 Higher education is devolved in the UK. The Department of Business, Innovation and Skills is responsible only for English universities and colleges. The administrations of Scotland, Wales and Northern Ireland have different policies on tuition fees. Scotland, for example, charges no fees to Scottish or EU students: those from the rest of the UK are charged £9000.
 The figure put on the national debt is obtained by adding up the nation’s liabilities – money owed – and subtracting the value of the assets owned. The net amount is ‘public sector net debt’. The government borrows to create student loans, but the loan is an asset, so you might expect the two transactions – borrowing and loan creation – to have zero net effect on PSND. But loans are viewed as ‘illiquid’ assets – they can’t be turned into cash easily – and are excluded from the headline statistics. This means that only the associated liability, government borrowing, is counted in PSND.