Skip to content

The cost of writing off tuition fee loans

DfE accounts published last week show that post-2012 loan debt had a face value of £46bn at the end of March 2017.
Tuition fee loans make up about 2/3rds of that. So abolishing post-2012 tuition fee loan debt would create a one-off hit of roughly £30bn to capital expenditure in the national accounts.
In the financial year 2016/17, £8.1bn of undergraduate tuition fee loans were issued. We can expect the debt write-off figure to increase by something like that every year, depending on future tuition fee rises.

Critical Education

The Green Party manifesto pledges to write off all outstanding student loan debt.  It estimates that this would mean foregoing £14billion in loan repaments over the course of the next parliament.

Unfortunately, serious consideration of this proposal has to sidestep the Greens’ own analysis, which is amateurish.

Firstly, the OBR figures cited by the Greens are for the whole of the UK, whereas HE is devolved issue and the UK government only has oversight of the English system. To measure the short-term cashflow impact on the national finances of abolishing English student loan debt (also includes tuition fee loans made to EU nationals studying at English universities) you need to subtract roughly £0.3bn per year to exclude the repayments being made by students from the rest of the UK. This brings the short-term cashflow figure down to about £13bn.

Secondly, you have to ignore the fact that they are…

View original post 864 more words


Fees frees from government control …

Jo Johnson’s speech and newspaper article at the end of last week succeeded in driving the news agenda back towards an idea first raised by IFS back in April: addressing existing student loan debt, particularly the tuition fee debt taken on by those who started at university in 2012 and after.

In the light of Labour’s pledge to abolish tuition fees, the latter group look like being hard done by. The cost of abolishing their tuition fee loan debt is much lower than the £100billion figure for the loans outstanding across the UK: assets worth £20bn would be erased with the cost appearing in the national accounts as a one-off £30bn hit to capital expenditure.

Johnson can consider his intervention to have been successful as it has distracted the media from his own travails over interest rates on student loans. As explained last week, the latest DfE accounts show that student loans went £500m over their “impairment” budget last year. Given that Justine Greening is also looking for large savings across the department, this might explain why Johnson was unable to indicate that he would be reviewing the interest rates due to come in this September (that 6.1% that’s cited everywhere).In fact, he made so bold as to tell his listeners to think about student loan repayments as a ‘graduate contribution’ and that anyway they wouldn’t find a better loan deal:

There are no commercial loans that offer this level of borrower protection. At 7.5%, the Bank of England’s reference rate for unsecured personal loans is materially higher.

Contempt for students has been a feature of the last decade in HE policy (both from politicians and senior university management). Johnson’s casual dismissal of concerns ignores the fact that much of the discussion in the broadsheet press has been dominated by the middle classes’ ability to remortgage (ie secured loans) and that alternative financial companies are offering CPI-only deals (but with administration fees and more onerous repayment terms). No defence of student loans can be premised on challenging students ‘to find better’.   It must be demonstrated to be a good deal for all borrowers, not just the credit-constrained. To that end, the government must discuss repayments frankly: at a minimum, ohnson  needs to require the Student Loans Company to restore the calculator, which was taken down last year after this website showed the flaws in its workings.

On a similar note, the DfE accounts might also provide further context for Johnson’s failure to announce the maximum tuition fees for 2018/19. It was expected that he would use Thursday to announce another inflation-indexed rise before Parliament closed for its summer recess. Perhaps the ‘value for money’ evoked in his speech’s title suggests a freeze is on the cards?

In which case, there would be added irony to one of Johnson’s key themes: how tuition fees ensure freedom from central government control.

Removing fees, and making the taxpayer bear the full cost of our higher education system, would have calamitous consequences for social mobility, it would be disastrous for teaching and research and it would be a hammer blow to taxpayers.


The immediate effect would be the reintroduction of strict student number controls, which would disproportionately reduce attendance by poorer young people with potential but without the grades.

… The second effect would be to starve universities of resources by making them compete head-on with all the other pressing claims on government spending, including our schools and our NHS. The experience of direct grant-funded HE systems, from Scotland to Continental Europe, is that they struggle to sustain per student funding.

In Scotland, the average Scottish-domiciled university place was underfunded by 6% in 2014/15, according to Audit Scotland. In Germany, which is debating reintroducing fees, per-student funding levels have fallen 10% since 2008. And we have our own experience of what can go wrong: in the two decades before England introduced fees, resource per student fell by over 40%.

I will finish this post with the following graph on “the unit of resource”. In his final sentence, is Johnson effectively saying, “Learn to love tuition fees, because they’ve spared you from the Conservatives”?

New Picture

Source: BIS, Research paper number 10: Review of Student Support Arrangements in Other Countries (September 2010), page 10.


Headlines from the DfE 2016/17 accounts

Updated 10.45am – I left out repayment data first thing.

Responsibility for HE (and student loans) passed from the department of Business, Innovation & Skills to the Department for Education last summer. I was led to believe that the DfE published its annual accounts much less promptly than BIS, but yesterday those for the year ending on 31 March 2017 appeared.

The headlines:

  • DfE has responsibility for £13.64billion loans issued last year.
  • There were £2.44bn in repayments made. Still very low – and a shortfall of over £11bn against new outlay. That contributes to Public Sector Net Cash Requirement which drives debt issuance.
  • The stock of debt – the ‘loan book’ – has grown to £89bn. That’s what you get from looking at all the oustanding balances – what is owed to government – and adding them together. This is known as the face value of the book.
  • The fair value of the loan book is now £61bn – this is what the government thinks the loans are worth. Here this is calculated by estimating all the future repayments generated by the existing loans and discounting them to a ‘present value’. The further in the future a repayment is, the less it is worth.
  • The impairment on new post-2012 loans issued  – ‘RAB charge’ – was £3.9bn. The government expects to get back much less than the equivalent of what it lent – £3.9bn less. As a percentage of new post-2012 loans issued this is 29% (£3.9bn/£13.575bn).
  • As I reported in February, the DfE applied for £11bn of additional resource to deal with changes to the key determinants for the value of existing loans (chiefly, Bank Base Rates and the rapid, recent upsurge in RPI (with determines the discounting measure – as this goes up, the value of future repayments is reduced).
    • DfE did not utilise all of this resource:
      • “Modelling improvements” wiped off £1.66bn from the value of existing loans;
      • Another £3.1bn came off as a result of changed “OBR forecasts”;
      • As a result, although there was over £11bn of net loan issuance in 2016/17, the fair value of the book  increased by under £4.5bn.

What’s the significance of all this?

The impairment on new loans issued is in the region of the 31% modelled for new starters in 2017/18 by IFS in its recent report. That DfE is just above its target RAB and RAB allocation for 2016/17 may mean that Jo Johnson has less room to manoeuvre around addressing concerns about interest rates applied to loans. He may need to find savings elsewhere or control loan outlay (for example by reviewing the commitment to increase tuition fees in line with inflation).   See target impairment correction.

Secondly, the volatility in loan value shows some of the difficulties with a sale – what the loans are worth to government are subject to fluctations. The loans earmarked for sale – pre-2012 loans – were adversely affected by the Bank of England’s decision to lower bank base rates to 0.25% last summer and by the upwards move in RPI. Interest on pre-2012 loans accrues at the lower of bank base rates plus 1 or RPI, so the interest rate is currently 1.25%. At the same time, RPI determines the fair value discount rate and increases the repayment threshold – in both cases lowering the value of the book (Repayments are lowered and the value of those future repayments is too!).
Pre-2012 loans went down in value by £2.9bn due those affects – those remaining have a face value of £42.8bn but a fair value of £29.7bn.  The government was hoping to raise around £12bn from a sale and its Value for Money uses a different discount rate, so that it is prepared to lose money on a sale. It will accept a price below that of the fair value – what it says the loans are worth.

The accounts mention the sale very briefly (with my emboldening):

Student loans sale

5.24 On 6 February 2017 the previous Government announced that the process to start selling part of the English student loan book had begun. To date, none of these loans have been sold. The new Government will take the decision about whether to continue with a sale process.

5.25 If a sale does proceed, the decision about value for money ahead of the sale would take account of a valuation of the loan book made on a different basis to that used to value the loans in the financial accounts. Under accounting policies, the amortised cost discount rate (currently 0.7%) applies in the financial accounts. Any decision to retain or sell an asset on the Government’s balance sheet involves an assessment of the retention value of the asset based on HMT’s Green Book principles where a discount rate must factor in a social time preference rate (currently 3.5%).

A higher discount rate means a lower valuation on future repayments – the Green Book rate (RPI + 3.5%) is much higher than the fair value discount rate (RPI +  0.7%).

Compare what’s in bold above to the bullish language in the 2015/16 BIS accounts: “The Department’s role in reducing debt will include the sale of the first tranche of the pre-Browne Income Contingent Student Loan Book.”

Some hesitancy from the government is good news here – selling the loan book for less than its worth makes no fiscal sense.

Fiscal Illusions

Today the Office for Budgetary Responsibility has published its biennial Fiscal Risks report. Outstanding student loan balances reached £100billion at the end of March with over 85% of those now the responsibility of the Department for Education. As such, they have an increasing relevance for thinking about the financial health of the nation.


OBR identifies various risks associated with student loans.

Firstly, they expect loans to grow at a rate of 0.8 per cent of GDP each year in the short-term, but should student numbers increase faster than estimates then loan outlay would have to increase and that requires more issuance of gilts. Loans are a national asset (money owed to government) but are created by borrowing, which is a national liability (money government owes).

Secondly, ‘economic drivers’ such as RPI, which affect interest rates, maintenance loan outlay and repayments. Graduate earnings are obviously another factor, with the OBR stressing that the spread of graduate earnings (not just averages or medians) matter in relation to income contingent repayment loans. That is, the number of graduates earning above the thresholds is crucial for repayments.

Thirdly, the sale of student loans. OBR describes further delays to the sale of student loans as the ‘largest downside risk’ in relation to the national balance sheet. I think there is some inconsistency here so I will quote from the report in detail.

§7.34 Student loan sales: These remain subject to market conditions and a final value-for-money assessment. Our March forecast assumed that the first sale would be completed in early 2017-18 and a second by the end of that year. These timings are likely to have been affected by the early general election. Selling the loan book affects the flow of cash to the Exchequer, with more recorded upfront as sales proceeds and less in future years, as repayments flow to the private sector instead. In effect this crystallises losses on the loans sold – the level of debt is permanently higher relative to no loans having been issued, because sale prices will reflect the interest rate and write-off subsidies implicit in the loans.

Elsewhere, OBR introduces adopts the IMF’s term ‘fiscal illusions’ to describe

… any transaction that improves or worsens measured fiscal aggregates without genuinely affecting the true health of the fiscal position in the same way. An example would be the effect of financial asset sales on PSND, where they lower the measured aggregate without improving fiscal sustainability. (§7.65)

It seems odd that OBR doesn’t flag up that student loan sales are exactly this kind of ‘fiscal illusion’ – they are financial asset sales – and moreover OBR doesn’t point out that the government’s value for money test for a loan sale uses the social time preference discount rate, a higher discount rate than that used to value student loans in the DfE accounts. Respectively, RPI + 3.5% (plus some adjusments) as compared to RPI + 0.7% – a higher discount rate puts a lower valuation on financial assets and so the government’s VfM test would sell at a loss.

It’s not just that a sale would ‘crystallises losses’ but it would worsen the fiscal position; the price may not ‘relect the interest rate and write-off subsidies’ but instead undervalue them.

Another fiscal illusion identified is the effect of student loan interest rates on public sector net borrowing (PSNB).

§7.13 Interest on student loans: this is recorded in PSNB as it accrues, which we expect to subtract £3.0 billion from the deficit this year. Interest starts accruing from the time the loan is extended and it is recorded within the public finances for the full amount owed rather than the amount expected to be paid. In reality some of this will never result in actual cash payments, because some borrowers will not earn enough to require their loans to be repaid. Eventually, this initial over-recording will be resolved by writing off any outstanding portion of the loan. But this may not be until years later – the write-offs associated with recently issued loans are not expected to pick up until the mid-2040s. So accruing interest will flatter the fiscal position in the meantime.

Although student loans as illiquid assets do not appear in PSND, the nominal interest accruing against their outstanding balances turns up as a positive in the deficit. At around £50billion in total, that £3billion is significant and may colour debates around the removal of real rates of interest on post-2012 loans.


IFS on tuition fees

As you’ve probably gathered from the media coverage, the Institute for Fiscal Studies has a new report out on tuition fees in England.

It offers a thorough summary of changes since 2012 and reinforces a number of points made here and elsewhere about the regressive nature of freezing the repayment threshold and how the abolition of maintenance grants plays out.

IFS even conclude with an oblique critique of the current system suggesting that government needs to increase its long-run contribution and be less focused on deficit reduction in this regard. In effect, IFS indicates that the cost of HE may now be too high and that this high cost ‘may reduce participation in the long-run’. This is awkward for Jo Johnson, who has spent the last few days using IFS findings to rebut Damien Green and the Labour pledge to abolish tuition fees.

The government is likely to review the interest rates, since these are scheduled to increase in September to a range of 3.1 to 6.1 per cent (based on March’s RPI figure of 3.1%). These figures were used by IFS in its analysis and underpin large increases in contributions made by the highest earners, such that the top-earning 30% of borrowers are now modelled to return a surplus of repayments to the government (the negative ‘RAB charge’ in the IFS chart below, where the resource accounting and budgeting charge represents the resource implications for the Department for Education, who now own the loans).

negative RAB

Politically a retreat from these high figures is probably in the government’s interest, but as IFS show, a reduction in interest rates would be regressive in that it would only benefit higher earners.

interest rate changes

(Note that the figures in this chart are ‘deflated’ to 2017 equivalents using CPI and so differ markedly from figures showing the cost to government based on the government’s discount rate of RPI + 0.7%).

IFS describe the high interest rates and ‘negative RABs’ as a risk to government, since if high earners find an alternative way to finance their study then their cross-subsidy is removed and the scheme becomes more expensive.

I’m not convinced the Treasury sees it so straightforwardly.

The government has in the past considered schemes to encourage early, voluntary repayment of loans (which would similarly allow individuals to escape the effects of real interest). In fact, current policy explicitly expresses its preference for cash today at the expense of long run loss: the sale of student loans is emblematic here.  If fewer people borrow in the first place or more repay early, then, all else being equal, public sector net debt is reduced in the short run.

The IFS insist that such measures do not improve the public finances – and they are right – but they improve key public finance statistics, like PSND, and politically this presentation matters. After all, what is the fiscal mandate other than an appeal to the public: judge us against these targets! and having PSND fall as a percentage of GDP is one of those targets.

On the flip side, but not noted by IFS, the Department for Education has been set a ‘target RAB’ of 28 per cent by the Treasury. For 2015/16, the RAB charge was well under that at 23%, but this was due to rise last financial year with the impact of maintenance changes; it will rise again in 2017/18 as higher tuition fees arrive. Should the official RAB charge rise above the target, then a special budgeting procedure kicks in which would see DfE having to make savings elsewhere in its budget. (DfE has a much bigger budget than BIS did, so the incentives behind the new budgeting procedure may have less effect now.)

The IFS have assessed the RAB charge on the 2017/18 cohort to be 31%.  This indicates that there may be pressure on the DfE budget, which as such would feature in discussions around changes to the interest rate. (We should note the official RAB is not calculated per cohort, but on each year’s loan issue and therefore covers students in more than one cohort).

For all the official reassurance around the English fee-loan regime, the number of tweaks to policy, loan terms and accounting / budgeting conventions is indicative of an experiment that is still in its early phases. It is also worth remembering that recent rapid rises in RPI will have caused significant downwards revisions to the value of existing loans and that last August’s decision by the Bank of England to lower base rates to 0.25% has similarly damaged the value of pre-2012 loans. In February, DfE bid for £11billion in supplementary resource to cover ‘movements in the macroeconomic determinants’ underpinning the value of the loan book’.

Further finagling of the system may be insufficient at a time when public goodwill seems to have moved decisively against the basic features of the scheme.







Teaching for 2017/18

I have updated my teaching page for 2017/18 courses open to the public at CityLit.

Creative arts & graduate earnings

I have a new piece at Wonkhe on last week’s Longitudinal Earnings Outcomes data and the creative arts.
It develops what has been said in my recent talks to Curating at Goldsmiths and to ILAS at Keele.