A beginner’s guide to student loans in the public accounts

Tesco is to pay out £235m to settle investigations by the Serious Fraud Office and Financial Conduct Authority … The settlements relate to Tesco admitting in 2014 that it had overstated profits by £326m. This overstatement was linked to how it booked payments from suppliers” – The Guardian, 2017

The way student loans appear in the public accounts – albeit entirely legal – has important similarities to the Tesco situation, so the report and blog by the Office for National Statistics (ONS) published just before Christmas proposing changes to fix the problem are very welcome.

The present system, with elements of a financing bubble, is unsustainable because it allows government to increase spending on higher education in a way that hides the losses on student loans for decades – or permanently. I discussed the issue more fully in my evidence to the Augar review.

The current loan system

Students can take out maintenance loans and fees loans. Graduates repay 9% of their income above a threshold (currently £25,000 per year). Loans charge an interest rate that varies from a zero real rate (i.e. a rate of interest equal to the rate of inflation) to a 3% real rate. Repayments start at the beginning of the tax year after the student leaves university, and any balance that has not been repaid 30 years later is forgiven. Thus someone who started a 3-year degree in 2015 graduated in July 2018, and will become liable to make loan repayments from 6 April 2019. Forgiveness applies to any balance outstanding on 5 April 2049.

Loans in the public accounts

The current system works as follows:

• The interest owed by past and current borrowers is treated as government current income, hence reduces the budget deficit (Public Sector Net Borrowing – PSNB). Thus the method counts as government income interest that is receivable this year but not necessarily paid this year, and may never be paid. The higher the interest rate, the greater the apparent benefit to PSNB.
• Write-offs are treated as a cost at the time they take place. A few, related to borrowers who die young, occur early, but most relate to forgiveness after 30 years. Thus write-offs for the cohort who has just graduated do not hit PSNB until 2049, and if the loan book is sold before 2049, never affect PSNB.

As a result the method inflates current income, and defers write-offs for 30+ years, flattering PSNB in the meantime.

The IMF describes accounting treatments that do not adequately reflect reality as ‘fiscal illusions’. These include any transaction that improves or worsens measured fiscal aggregates without genuinely affecting the true health of the fiscal position in the same way” – Office for Budgetary Responsibility, 2017.

None of these findings is new.

Bluntly, the main motive for replacing T grant by loans is an accounting trick. There is an apparent decline in public spending, but at the cost of distorting higher education policy … Thus the changes look like a dodgy [Private] Finance Initiative” – Barr, 2012

Even Private Eye, not widely known for expertise in public accounting, had rumbled the system:

“With … experts predicting that, given the repayment terms and defaults, the new system will be no cheaper for taxpayers, who will benefit? The short answer is the government, whose spending figures will be flattered in another huge accounting fiddle reminiscent of PFI … the costs and debt from funding higher education have magically disappeared from the books for several years – and most importantly until well after the next election.” – Private Eye 1291, June 2011

How did this happen?

With a bank loan or mortgage – fixed monthly repayments for a fixed duration and a relatively small expected loss – it makes sense to count interest accruals as income, and to book losses when they occur. Student loans, however, are different. Forgiveness after 25 or 30 years means that they make a loss on borrowers with low lifetime incomes by design. When student loans were small with a low interest rate, postponing booking those losses was still technically wrong – but since the effect was small did not merit attention.

The 2012 reforms changed matters in three ways. The maximum loan was increased and so was the interest rate, both of which increased measured government income; and the repayment threshold was increased in 2012 and again in 2018, increasing eventual write-offs. Current estimates suggest that the loss on student loans (the so-called RAB charge) is about 45% of total lending. This is very different from a conventional bank loan.

There are good reasons why government can appropriately behave differently from private companies, but the 2012 reforms tested the system to destruction.

Why does it matter?

Why should anyone care about any of this? Answer: because the accounting method distorts policy for the worse.

The starting point is to consider how much subsidy for student loans is desirable. The purpose of these loans is to allow young people access to their own future earnings – what economists call consumption smoothing. However, borrowing to finance the acquisition of skills is risky because (in contrast with home loans) there is no collateral. As a result, people will invest too little in acquiring skills. To avoid that problem, a well-designed loan has inbuilt insurance.

The income-contingent formula provides insurance against low current income, and forgiveness after 30 years against low lifetime income. The subsidy to people with low lifetime earnings – the insurance element in the loan – is well-targeted social policy spending.

Subsidy beyond that – notably the sharp increase in projected non-repayment after 2012 – is not well targeted. As stated in a report by the House of Lords Economic Affairs Committee: “Recent changes to higher education financing have been motivated mainly by the desire to lower the deficit”.

Specifically, the ill-effects of a system include:

• A bias towards loans: a system in which loans are apparently costless in PSNB terms create a bias towards loans and hence militate against taxpayer finance for teaching (T grant) and maintenance grants, and against spending on pro-access policies earlier in the system.
• A bias towards 3-year degrees: apparently costless loans mean that higher education is better financed than level 4 and 5 technical qualifications. As a result higher education is better financed than further education.
• Adverse distributional effects: the large losses on student loans benefit those who have made it to university, rather than using the resources in further education and for activities earlier in the system. The result is to benefit insiders as the expense of outsiders.
• An unstable basis: because higher education finance has elements of a bubble. If I were a Vice-Chancellor, this aspect would give me sleepless nights.

Nor do the problems end there.

Maintenance loans are too small and include parental contributions, harming access; loans for part-time students are inadequate, also harming access; loans for postgraduate students are inadequate; and there are virtually no loans for other parts of tertiary education including vocational education.

In sum, while there is ample room for debate about policy design, the downward bias in the measured deficit is unambiguously damaging. Education policy should be determined by economic analysis not statistical artefacts. This is not the way to run a railroad.

The proposed change

The necessary policy change is that the projected loss on student loans issued this year should score as public spending this year.

The Government is not responsible for the international accounting rules that allow the fiscal illusions within student loans to exist. However, the National Accounts accounting rules regarding financial transactions were not intended to be used for loans that, as the Government readily promotes, are designed to not be paid back in full.

Loans that are intended to be written off are, in substance, a partially repayable grant rather than a loan. The ONS should re-examine its classification of student loans as financial assets … and consider whether a portion of the loan should … be classed as a grant” – Treasury Committee, 2018

Most student loans will not be repaid in full: some will be paid in full, some not at all, and a lot only partially repaid. The expected write-offs should be shown in the deficit when the loan is issued. The true cost of funding higher education would then be immediately apparent” – House of Lords Economic Affairs Committee, 2018

That change is what the ONS report recommends.

We have decided that the best way to reflect student loans … is to treat part as financial assets (loans), since some portion will be repaid, and part as government expenditure (capital transfers), since some will not…

“This decision means that the impact of student loans on public sector net borrowing … will better reflect government’s financial position. This is because government revenue will no longer include interest accrued that will never be paid; and government expenditure related to cancellation of student loans will be accounted for in the periods that loans are issued rather than at maturity”

Pros and cons

No reform is perfect. Under the proposal, the projected loss on loans issued this year counts as public spending this year. But a projection is just that – a projection. Such projections are complex, and estimates will need to be adjusted over time. But this is a case where it is better to be approximately right than exactly wrong – especially where “exactly wrong” creates powerful incentives to bad behaviour by government.

The major gain from the proposed change is that it aligns the accounting of loans with their resource implications and thus reverses the adverse effects discussed earlier.

The change removes the artificial bias towards loan finance. If the government makes repayment terms more generous (e.g. the increase in the repayment threshold from £21,000 to £25,000 in 2018), the cost will affect the budget deficit (PSNB) this year, not in 30+ years time, thus fundamentally changing political incentives.

There is no longer an artificial bias against taxpayer finance. At present, a £1,000 loan, by inflating government income, makes PSNB smaller, while a £1,000 grant increases PSNB by £1,000. Under the new arrangement, if the marginal RAB charge is 50%, a £1,000 loan increases PSNB by £500, and £1,000 grant by £1,000. Thus the relative price of restoring T grant and/or maintenance grants has halved.

Analogously, the new arrangement creates a more level playing field between higher education (largely loan financed) and further education (mainly taxpayer financed). And if the improved political incentives encourage a less costly loan system, the regressive element in loan design is reduced.

The new arrangement means that HE finance is no longer partly dependent on a bubble. The longer the bubble is allowed to go on, the bigger the bust at the end. Getting it right now places university finance on a sounder footing. In this respect, at least, vice chancellors should sleep more easily.