The Silence of the Loans: How student loan book fictions prevent higher education reform

The top line

The government has already borrowed the money it spends on higher education. The student loan system does not replace public funding with private – it disguises public spending through an accounting mechanism that treats income-contingent, time-limited, unenforceable repayments as though they were a financial asset equivalent to a commercial loan. This fiction, now valued at around £300bn, blocks every serious option for reform.

What should be done

The Treasury, OBR, and ONS should work together to reclassify the loan book, model alternatives on a like-for-like basis, and create fiscal rules that permit revenue-neutral restructuring. The moratorium on loan book sales should be extended. Annual reporting should make visible the divergent trajectories of the Plan 2 and Plan 5 books. These steps do not choose a replacement funding model – they make rational choice possible.

The bottom line

The accounting treatment that once made higher education funding disappear from the public accounts now makes higher education reform disappear from the policy agenda. The question is not whether the government can afford to reform the loan book – the money has already been spent. The question is whether it will permit an accounting convention to forever determine the boundaries of the possible.

Contrary to what is often asserted, the student loan system, and in particular the 2012 changes introduced by the Coalition government, did not significantly reduce the costs to the state of higher education nor replace public subsidy with private money. Rather, the changes traded transparent, accountable, and intentional public grant funding for money still raised by the state, but now through a mechanism that was opaque, poorly monitored, and highly volatile.

The effect of this fiscal sleight of hand has been to unnecessarily block discussion of how to do almost anything necessary for the English HE system to thrive. Whether policy-makers wish to reduce the perceptions of unfairness experienced by new students and recent graduates, ensure all students receive a high quality education, or ensure greater value for money for the taxpayer from the system, they find themselves stymied by the argument that the cost is too great because reforming the loan system is too expensive.

But this is, bluntly, no more than sophistry: the government has already borrowed the money it pays out to higher education, and the government is as – if not more – capable of repaying that borrowing through general or graduate-specific taxation as it is through student loan repayments. One system of accounting for higher education spending has been accepted as more favourable than others through what is frankly an accounting trick and not a genuine reckoning of the value of these policies to students, institutions, taxpayers or the wider economy.

To understand how the trick is done, it is necessary to appreciate a simple distinction in the way the government keeps its accounts, the difference between grants and loans.

When a government spends money as a grant, it appears, perfectly logically, in the public accounts as expenditure. In times of reduced government income, such grants might also widen the deficit – the gap between what the government receives and what it spends – if the government found it necessary to borrow to make the grant. What the grant was spent on is irrelevant: even if the government is very confident that a grant is likely to generate economic returns that eventually increase tax revenue, because those returns are uncertain, that revenue does not appear in the accounts until it materialises. As such, grant expenditure appears on the spending side of the government’s accounts only.

When a government lends money, something different happens. The cash goes out, yes, but the government has acquired something in return: a contractual right to future repayments. That right is a financial asset and thus has its own value, and because the government has simply exchanged one asset (cash) for another (the loan), the deficit need not move at all. If the government borrowed the money it then lent out, that borrowing is recorded, and the national debt (the total amount the government owes) increases accordingly – but the deficit is unaffected.

The cause of this asymmetry is right to repayment, irrespective of the circumstances of the borrower. With a loan, in the event of missed payments or complete refusal to repay, lenders have the right to pursue legal action and clear mechanisms by which they can do so; even if the borrower is genuinely unable to meet the loan’s terms because of the poor state of their finances, they can be declared bankrupt, and the lender can expect to recover at least some of the value of the loan.

In contrast, although a government can compel citizens to pay tax, it cannot compel citizens to undertake taxable activity. Thus, even if a grant is highly likely to result in behaviour that will increase tax income, there is no asset on the balance sheets, because individuals and institutions may choose not to undertake the desired taxable actions, and the government has no legal mechanism to force them to do so.

This distinction makes perfect sense in the case of the forms of lending most people are engaged with, such as car loans, credit cards, or mortgages, in which the borrower is required to repay in full, in pre-determined sums, to a pre-determined timetable, and can be legally sanctioned in the event they do not keep to those terms.

But anyone conversant with the terms of England’s income-contingent repayment student loans will immediately realise that those terms are materially different to traditional lending. Repayment depends not on the sum borrowed but on what the borrower earns, and ceases entirely below an earnings threshold, and the loan also gets written off after a period of time, irrespective of how much remains to be paid. Therefore, the government cannot compel repayment if the borrower is not working, and cannot force a borrower to work in order to pay off the loan.

Demonstrably, then, England’s student loans operate in a manner identical to taxation, and not like a loan at all. The claim that the loan book is an asset rests on misleading semantics: because the government has funnelled the money through a thing it has called a loan, the expenditure is balanced in the national accounts with a corresponding asset. But student loans lack the specific characteristic that justifies treating loans as assets – borrowers are not required to pay in full, cannot be compelled to pay if they are not working, and the loan is written off after a fixed period regardless of what remains.

Such creative accountancy matters because it distorts any discussion about English higher education policy by seemingly perpetually embedding an immovable block at its centre: the loan book asset, currently valued at around £300bn. A government might rationally wish to replace loans with grants, reduce fees, restructure repayment terms to be less burdensome on lower-earning graduates, write off some portion of the debt, or move to a graduate contribution collected through the tax system. But every one of these options involves reducing the expected value of the loan book, and every reduction in the expected value of the loan book is, in accounting terms, a loss – a “destruction” of a financial asset that hits the deficit and worsens the debt-to-GDP ratio.

But that £300bn asset is not real: it exists only because the government has asserted that student loans are entitled to be treated in the government’s books as other types of loan, despite lacking the very legal enforcement mechanisms that justify creating counterbalancing assets when loans are made.

Advocates of the loan system might retort that, even if the student loan book does not share the enforcement mechanisms that other loans have, its claim to be an asset can be validated because, like any other asset, it can be sold on to others. But when the government did sell parts of the income-contingent repayment loan book, in 2017 and 2018, those sales demonstrated only that the government cannot properly assess the value of the asset: the OBR found that significantly less had been raised than planned, the Public Accounts Committee determined that even the desired sale price would have been too little to justify the transaction, and the ONS subsequently redefined how student loan sales could be recorded in accounts, so the whole thing was called off in perpetuity anyway.

Aspects of the government’s loan legerdemain have been recognised before. Between 2012 and 2018, the full face value of each loan was booked as an asset, and the interest accruing on loans – including interest accruing on sums that would never be repaid – actually improved the government’s fiscal position on paper, because it increased the value of the asset. But the expected losses would not hit the deficit until the loans were written off, three decades later.

The Office for Budget Responsibility called this a “fiscal illusion”, which is a restrained way of describing an accounting treatment in which the government could lend billions of pounds a year, fully expecting to lose nearly half of it, and record no cost whatsoever. Having consulted with international statistical bodies – principally through the European System of Accounts 2010 (ESA 2010) framework, which required a more honest treatment of government-issued loan portfolios – the ONS announced that it would partition student loans into two elements: the portion the government could reasonably expect to recover, which would continue to be treated as a financial asset, and the portion it would not see again, which would henceforth be recorded as government expenditure at the point of issue rather than at the point of write-off thirty years later. Thus, the deficit could no longer be flattered by phantom interest on sums that were never coming back and the subsidy embedded in each year’s lending was made visible immediately – but the fiction that the remainder was an asset legally enforceable in the same way as other loans was retained.

It matters that we are clear about this, because much discussion of viable future options for English higher education accepts that 2012 marked an irreversible watershed, in which private money substituted for public, and that any attempt to return to the status quo ante would be unimaginably expensive. But in fact the amount of money the government spent on higher education in 2010, 2012, and 2015 was broadly similar. Instead of grant funding paid to institutions through the Higher Education Funding Council for England (HEFCE), roughly the same amount instead flowed through the Student Loans Company (SLC) to providers, in accordance with the number of students they had.

This shift was not purely about accountancy; it operationalised a view of the world in which funding routed through individual student choice would drive greater efficiency, improve quality, and improve labour market impacts of HE spend. On this view, the value of the loan mechanism is not the asset it creates on the balance sheet but the steering mechanism it embeds – one that privileges student choice over Whitehall direction. None of those things happened, for reasons I have explored elsewhere.

And because the SLC is wholly owned by the British state, in terms of where the money actually came from, this changed nothing anyway. Money that the Treasury had previously given to HEFCE to distribute according to its policies was instead given to the SLC to distribute according to its own – that those policies were quite distinct (HEFCE gave grants to universities; the SLC loaned money to students) did not change the source of the funding. Indeed, the SLC tuition fee money resembled HEFCE’s funding more closely than the above description suggests, since the student borrower never saw it: it went directly to whichever provider the student had nominated by accepting a place on their course.

Ultimately, the thing that changed in 2012 was not a reduction in public funding, but (temporarily) an increase: the near-universal adoption of the £9,000 fee – which the Coalition had confidently predicted most providers would not charge – meant the government was underwriting substantially more lending than it had anticipated. For a time, the fiscal illusion shielded this from political scrutiny, but the underlying reality was that the state was spending more on higher education than before, not less, while appearing to spend nothing at all.

So, this is not for a moment to suggest that nothing important changed in 2012, but it was not a matter of the introduction of new sources of income in the system. Rather, England replaced a system where most of the money spent on higher education by the state was in the form of a grant whose workings were matters of public record, the purposes of which were publicly avowed by an elected government, and which could be influenced by agents of the state to achieve desired policy ends, to a system in which, despite the money flowing through a government-owned company, it has become impossible to meaningfully influence the direction of the flow through government action, breeds hyper-bureaucracy to hold anyone in power accountable for the consequences of the spend, and does not even get very far in finding out on what, and with what result, higher education institutions spend this income.

What did change in 2012 was how the overwhelming majority of public funding for higher education appeared in the government’s accounts, and, to the casual observer and until 2018 at least, it more or less erased the costs of higher education from view entirely.

Once the 2018 reclassification forced an honest accounting of at least part of this alleged asset, the government sought to obscure the issue once again: from the 2023/24 academic year, new borrowers in England were placed on Plan 5 – a loan with a lower repayment threshold of £25,000 (frozen until 2027, then rising with RPI), a longer repayment window of forty years rather than thirty, and interest set at RPI only, with no additional margin during study or afterwards. As a result, the government could once again claim it had a valuable asset which now would require much less of a write-down at the end of the longer loan term. The RAB charge for Plan 2 full-time loans has fluctuated dramatically as successive governments changed the terms: from above 40 per cent in earlier forecasts to 32 per cent in the latest DfE figures, with independent modelling by London Economics suggesting the true figure for the 2022 cohort may now be negative. Under Plan 5, the DfE projects 29 per cent. But these numbers are artefacts of the accounting framework the paper has been describing — they move not because the system has become cheaper, but because the terms have been tightened on borrowers.

Ministers in the previous parliament sold this explicitly as putting student finance on a more sustainable footing by ensuring more students would repay in full, but their treatment did not alter the fundamental issue that the government does not, in any meaningful sense, possess a right to demand repayment of these loans in any way that is distinct from the state’s right to claim tax from citizens. That Treasury has changed the loan terms to increase the amount of money they will take from graduates’ pay-packets and the length of time for which they will do changes nothing about the enforceability of these loans in the event of un- or under-employment of graduates. Plan 5 is as much as obvious tax analogy as Plan 2 was.

In fact, Plan 5 is actively worse for students and taxpayers: it is distributionally regressive because the highest-earning graduates – those most able to bear the cost – repay quickly and escape the system entirely, paying less in real terms than under Plan 2. Middle earners, by contrast, now face forty years of repayment at a lower threshold, paying more over their lifetimes than they would have under the previous terms. The graduates who will carry the heaviest burden are precisely those for whom the system was supposed to work: professionals in teaching, nursing, social work, and the public services on which the country depends. This is not an incidental feature of the design, it is the mechanism by which the RAB charge was reduced – and it guarantees that the political anger which has already erupted among Plan 2 borrowers will, in due course, be repeated (with, if you’ll forgive the pun, interest).

Because it continues the asset fiction, Plan 5 does not render an honest accounting of England higher education spend – the assumption that Plan 5 will return more money to the Treasury is predicated on an almost heroically-optimistic projection of the likely long-term earnings of graduates. But it does impose a greater burden on graduates, to which they are likely to react with fury as they progress through the workforce and the scale and consequences of their debt becomes clearer to them. Yet none of the Plan 5 changes address the underlying structural questions about what the money is for, how it reaches institutions, or whether the incentives it creates serve the public interest.

All Plan 5 has done is bake into the government’s accounts a version of the current system which will undoubtedly be as, if not more, unpopular in due course, without improving the government’s capacity to achieve anything with the money it is spending on HE. The Plan 2 book – with its higher RAB charge, its thirty-year write-off clock already ticking, and its more generous terms for borrowers – will continue to dominate the balance sheet for decades, the first significant Plan 2 write-offs begin in the early 2040s and the fiscal politics of that moment have not been addressed, merely deferred. But, as a new Plan 5 book accumulates alongside the old one, the total value of the loan asset on the government’s balance sheet will continue to grow, deepening the perception trap that prevents reform.

The damage of Plan 5 goes beyond that which it does to the graduates who will repay on its terms: the continuing conviction that these student loans can and should be described as an asset is actively damaging the capacity to reform the system, for fear of how such reforms will be presented, irrespective of the underlying fiscal reality.

Under Plan 5, the performing portion – the repayments the government genuinely expects to collect – remains a very large financial asset on the public balance sheet. And with Plan 5 cohorts entering repayment from April 2026, that performing portion is projected to grow as a share of each year’s lending. The bigger the portion gets, the bigger the seeming “destruction” of the asset is if anything is reformed.

This is true even when the proposed alternative would leave the government’s actual fiscal position unchanged or improved. Suppose, for example, that the government wrote off the entire loan book but simultaneously introduced a graduate levy, collected by HMRC through the payroll system above an earnings threshold – mechanically identical to the current repayment system, potentially yielding the same or greater revenue, but structured as a tax rather than a debt repayment. The government’s cash flow would be unaffected. Its annual revenue would be maintained; indeed, over the long term, it might even collect more, because a permanent tax has no thirty- or forty-year expiry date.

But, as currently rendered, the public accounts would show a catastrophe. The loan book – the financial asset – would be destroyed. And the replacement revenue, however certain, however permanent, however large, cannot be recorded as an asset.

The result is an absurdity with profound consequences. The government’s balance sheet recognises a depreciating, demonstrably fictitious, time-limited financial asset as being worth tens of billions of pounds. It would not recognise a permanent, enforceable, HMRC-collected revenue stream as being worth anything at all until each pound arrived. A reform that improved the government’s actual fiscal position would, on paper, look like the largest voluntary balance sheet event in modern British fiscal history. The debt-to-GDP ratio would spike. The fiscal rules – whatever their current incarnation – would be breached. And all this despite the fact that the reasons to treat loans differently from taxation do not apply in this case anyway!

This is the trap. Not a trap of fiscal reality, but a trap of fiscal representation. The accounting treatment of the loan book makes the status quo look cheaper than it is – a wasting asset, deflating year by year as cohorts reach write-off, valued on assumptions about graduate earnings and participation rates that the demographic outlook is already invalidating – and makes every alternative look more expensive than it would be. The system is locked not by economics but by accountancy. Plan 5 tightens the lock, because its lower RAB charge makes the performing asset look healthier, even as the underlying system remains incapable of directing public resources toward public ends.

Before examining what the accounting trap prevents, it is worth stating plainly what a funding system for higher education ought to accomplish. At minimum, three things:

• first, sustainable and dependable funding for a sector vital to the common good, including to the economy (through both its research and its teaching functions) and to democratic society (through its creation, curation, and communication of knowledge as a critique and mirror of that society, and through the socialisation of a significant portion of young adults);
• second, that whatever aspect of that funding is public – however raised – should be directable toward public purposes, so that the state retains the capacity to appropriately influence the shape, quality, access, provision, and the relationship between higher education and the economy;
• and third, that those who most personally benefit from higher education should (and should be seen to) contribute a fair share of the cost of their education, set at a level that is politically sustainable across generations and does not deter participation.

If the trap merely preserved an imperfect but functional funding system, it might be tolerable. But the cumulative evidence of the past 30 years is that the English system is progressively less capable of applying public resources to educational quality improvement, systemic financial stability, or sustained confidence in the value and outcomes of higher education amongst students and taxpayers.

Although grant funding does not guarantee improved policy making, it does create circumstances in which policy instruments can more intentionally distribute funding. Current government grant funding now operates on a scale dwarfed by the funds pushed through the loan mechanism, and therefore grant funding incentives do not outweigh the cumulative effect of aspirant students’ choices. But the loan book militates against the restoration of grant funding, because grants are pure expenditure with no offsetting asset creation and therefore look more expensive, even when the loan they are replacing was never going to be repaid. It closes down discussions of meaningful reform of repayment terms, because any change that reduces expected repayments triggers a write-down of the asset. It makes fee reduction look ruinously expensive, because lower fees mean lower loans which mean a smaller asset being created each year.

If, as I have argued elsewhere, the student choice-led model has not, in practice, created the responsive, self-correcting market its designers envisioned, the only remaining credible argument for the current system is that it provides a guarantee of the autonomy from the state which the function of universities especially requires. Linking recruitment and income means institutional funding depends on ability to attract students, not on the priorities of whichever minister happened to hold the brief. Grant funding or a hypothecated tax, it is argued, would leave institutions vulnerable to political will, or whim.

But this is as much an illusion as the accountancy trap. The current system has not protected universities from political interference – it has simply allowed government to disclaim responsibility for the consequences of a system it designed and funds. The result is the worst of both worlds: institutions are neither genuinely autonomous (the interventionary apparatus has grown, not shrunk, since 2012, including through a Home Office conspicuously unbounded by respect for sector autonomy) nor genuinely accountable to a public that has no visible stake in how they are funded.

If the public had a more direct and transparent role in funding higher education, both institutions and government would face greater pressure to demonstrate that the system works as intended – and that pressure, uncomfortable as it might be, is precisely what has been missing.

Yet, for much of the political class, the current accounting arrangement forestalls the most structurally rational alternatives – a graduate contribution through the tax system, a return to formula funding, any hybrid that breaks the contractual link between individual borrower and individual debt – because all of them involve “destroying” the asset, even though the underlying reality of income expenditure may remain unchanged or even benefit the Treasury, because the accounting framework has no mechanism for recognising such a shift.

It is notable that even the Augar review – the most serious official attempt to address the funding question – accepted the trap’s premises. Its recommendation to reduce the fee cap to £7,500 and partially restore teaching grants was costed within the existing accounting framework, meaning the fiscal impact was calculated by reference to the effect on the loan book asset. The result was that a policy which would have restored some grant directability was presented as requiring significant additional public expenditure, when in reality the quantum of public money flowing through the system would have changed far less than the headline figures suggested. Augar demonstrated, inadvertently, that no reform costed within the current accounting architecture can escape the trap.

But, so long as this new fiscal illusion is accepted, no meaningful funding reform can take place, and worse, the perceived impediments grow, not shrink, over time. The Plan 2 book is a wasting asset with a defined lifespan – a closed cohort, no new borrowers, declining to nothing by the early 2050s, but the Plan 5 book is not. Because new cohorts enter every year, the Plan 5 asset replenishes itself indefinitely: older cohorts move through the forty-year repayment cycle while new lending restocks the book, and the Plan 5 asset does not waste to zero, instead reaching a permanent steady state, meaning the trap described will persist, even as the political consequences of Plan 5 loan construction become ever more unpalatable for debtors and thus the governments they elect.

In short, the trap binds closer and closer the longer we are caught in it.

The question should not be “can government afford to reform the loan book”, but instead, “should government permit itself to be bound in perpetuity by an accounting trick?”.

The only way out of the trap is to expose the falsehood at its heart, to ensure that the money government has already borrowed to fund higher education is properly recorded in the national accounts, and to level with the public and the bond markets that the deficit and the debt-to-GDP ratio have been mis-stated, but that nothing material has actually changed.

Such measures have been undertaken previously. After the 2018 reclassification, the OBR scored the change as adding roughly £12 billion a year to the measured deficit – not because the government was spending more, but because it was finally admitting to spending it had been doing all along, but significantly, this had little to no effect on the cost of government borrowing or the public perceptions of the government’s financial competence. The order of magnitude involved in writing off the loan book asset is obviously greater, but such exceptional accounting events occurred during the Global Financial Crisis (GFC) and while by no means routine, they clearly can be done – and, in this case, have the advantage that they do not create new borrowing but simply represent properly in the national accounts the fiscal reality.

Below, I make six recommendations, designed in sequence. The first two create the fiscal architecture that would make reclassification technically possible and politically defensible. The third forces into existence the like-for-like comparison which we currently lack, and which flatters the current system. The remaining three are protective: they prevent the government from selling off optionality while reform is under discussion, create fiscal rules that distinguish restructuring from profligacy, and make the cost of inaction visible year by year.

I do not make any recommendations here about what system of revenue-raising ought to replace the present one, not for want of ideas, but because as my previous work argued, that answer needs to be generated in a coherent manner alongside other changes to the system’s control factors. These recommendations dismantle the apparatus that currently prevents any rational choice from being made.

The following recommendations are directed not at choosing a specific alternative funding model for English higher education, but at dismantling the accounting architecture that currently prevents any rational choice between models from being made.

1. OBR balance sheet reclassification model.

The OBR should be commissioned to model a one-off reclassification of the student loan book, analogous to the treatment of bank nationalisations during the GFC. Those purchases were treated as one-off capital events excluded from headline deficit measures. The OBR should set out the conditions under which a loan book restructuring could be similarly classified.

2. Treasury/ONS review of hypothecated revenue accounting.

HM Treasury and the ONS should examine whether a legislated, permanent, HMRC-collected graduate contribution can be recognised as a fiscal offset to loan book write-down. The current asymmetry – in which a depreciating loan asset is valued at tens of billions while a permanent revenue stream is valued at zero until each pound arrives – is a feature of an accounting framework designed for a world of commercial lending. It requires the accounting infrastructure to stop making that choice look artificially expensive.

3. OBR comparative costing report.

The OBR should produce a standalone report scoring the lifetime fiscal cost of the current system against at least three alternatives: a graduate contribution collected through HMRC, a return to formula funding with reduced fees, and a hybrid model. No such like-for-like comparison currently exists. The comparison should be conducted over a fifty-year horizon, capturing the full write-off cycle for both Plan 2 and Plan 5 cohorts, and should incorporate the demographic projections already calling into question the loan book’s valuation.

4. Extend the moratorium on loan book sales.

Every tranche sold permanently removes optionality. Sales should be explicitly suspended pending completion of the comparative report. The government has already sold tranches of the pre-2012 book at prices the ONS subsequently identified as significantly below carrying value. Selling while the system is under review is to dispose of an asset whose value depends on policy decisions not yet made.

5. Fiscal rules provision for revenue-offset restructuring.

A restructuring accompanied by a legislated replacement revenue stream of equivalent or greater value, as certified by the OBR, should not count against debt or deficit targets. Fiscal rules exist to prevent irresponsibility; they should not also prevent rationality.

6. Annual dual-book transparency reporting.

The Department for Education should be required to publish annually the divergent trajectories of the Plan 2 and Plan 5 books: the year-by-year decline of the closed Plan 2 cohort to zero, the cost of Plan 2 write-offs as they materialise from the early 2040s, and the steady-state growth of the open-ended Plan 5 book. This projection should make visible that the government will hold a permanent financial asset on its balance sheet that permanently constrains reform options – not as an interim cost of transition, but as an enduring feature of the fiscal architecture.

The student loan book is not an asset. It is a political fiction that has outlived whatever usefulness it once possessed. For over a decade, it shielded the true cost of higher education from public view; now, perversely, it shields the true cost of inaction. Every year that the fiction is maintained, the Plan 5 book grows, the trap binds tighter, and the political consequences compound.

The question facing any incoming government is not whether it can afford to reform the loan book – the money has already been spent. The question is whether it will permit an accounting convention, designed for a world of commercial lending that bears no resemblance to income-contingent student finance, to forever determine the boundaries of the possible for English higher education. Until that convention is broken, no coherent post-18 system can be built.