Education

Inflation is papering over a £6.4bn hole in the student loan book

Inflation is papering over a £6.4bn hole in the student loan book


Four months late, the Department for Education’s accounts for 2025–26 say that England’s student loans lost just £30m in value last year.

That sounds like good news, until you look at the rest of the figures.

Student loan accounting has a reputation for being impenetrable, and some of that is deserved. But the ideas underneath it are simple enough, so I’ve tried to take them one at a time.

I’ve looked at what the government says graduates owe, what it actually expects to get back, how it decides, and why this year’s headline figure hides some much bigger ones with some fairly miserable implications.

Just on the lateness. Government departments are expected to publish their accounts before Parliament breaks up for the summer, which this year meant 16 July. DfE missed that, along with three other departments, and in July the National Audit Office (NAO) – the body that checks government accounts – would only say that it was still working through “matters identified” in its audit, and expected to finish in October.

It actually signed off on 12 November. That’s within the legal deadline of 30 November, but there’s no explanation of what took so long – and as we’ll see, most of the clues in the accounts point at student loans.

Owe dear

Imagine you’ve lent money to lots of friends. Add up all the IOUs and they come to £100. But you know some of them will never be able to pay you back, so if someone asked you what those IOUs are really worth, you might say £57.

That’s what DfE does. Add up every English student loan statement and graduates owe a total of £293.2bn. But on its own books, DfE says those loans are worth £165.8bn. The difference – £127.4bn – is what it doesn’t expect to collect, mostly because whatever is left on a graduate’s balance gets wiped after 30 or 40 years.

So for every £1 on a statement, DfE now expects to get back about 57p. A year ago it was 60p. In 2023 it was 68p. Ouch.

Over the year, what graduates owe went up by £28.5bn, but what DfE thinks it will get back went up by only £7.9bn. Almost three quarters of the new debt that appeared on graduates’ statements is money the government has already accepted it won’t see.

Most of that is interest. Take Plan 2 – the loans for anyone who started university between 2012 and 2022. No new students have joined it for three years, yet what they owe still grew by £11.6bn last year, because £11.3bn of interest was added to balances while graduates repaid £3.3bn. DfE expects to get back 56p in the pound on Plan 2, and it makes up £99bn of the £127bn gap.

The newer loans for English undergraduates – Plan 5, for those who started from 2023 – were designed to cost the taxpayer less. But in just one year, what DfE expects to get back on them fell from 78p in the pound to 66p.

Loss leaders

Every time DfE lends a student money, it immediately notes down the part it doesn’t expect to come back as a cost – on the day the money goes out, decades before anything is actually written off.

You’ll sometimes see this called the RAB charge. In practice, it’s the clearest measure of how much the taxpayer is putting into the loan system each year.

This year DfE lent £21.2bn, and straight away wrote off £7bn of it – 33p in every pound, up from 30p last year. For the Plan 5 loans it was 33p, and for the Plan 2 money still going to students finishing their courses it was 39p – the same figures DfE’s summer forecasts had predicted.

Chart showing the share of new student lending written off on day one, from 52 per cent in 2020-21 to 33 per cent in 2025-26

So why has it gone up? To answer that you need one more idea.

A pound you’ll get back in 30 years’ time isn’t worth as much as a pound in your hand today – prices will have risen, and you could have used the money in the meantime. So when the government works out what future repayments are worth, it shrinks them down to “today’s money”. How much it shrinks them depends on a rate that the Treasury sets each year, based on what it costs the government to borrow.

This year the Treasury nudged that rate up. A higher rate shrinks future repayments more, so the loans look worth less and more gets written off on day one – even though nothing about graduates has changed. That explains a lot of this year’s rise.

Take off the profit – yes, profit – that DfE makes on master’s loans, and the taxpayer’s contribution comes to about £6.9bn for the year. The Office for Budget Responsibility (OBR), which does the official forecasts for the government, puts it at about £7.6bn using a slightly different method – something we’ve looked at before.

Paper over the cracks

Now to that £30m. As well as writing down new loans, each year DfE re-guesses what all the loans it has already made are worth. It does that by re-running a computer model of what millions of graduates will earn over the next 30 or 40 years, and so how much they’ll pay back.

This year the model concluded that graduates will earn less than it thought a year ago – and so pay back less. That knocked £6.4bn off the value of the loans. Last year the same thing knocked off £6.7bn. And this time it’s the newer loans too – Plan 2 lost £4bn, and Plan 5 lost £2.1bn, when last year its value went up.

So how does a £6.4bn hit end up as a £30m loss? Go back to shrinking future money into today’s money. Every year that passes, all those future repayments get a year closer – so they get shrunk a little less, and the loans look worth a bit more, without anybody paying a penny.

That “time passing” gain gets bigger when inflation and the Treasury’s rate go up, and this year they did. It was worth £5.6bn, up from £4bn last year. The change in the Treasury’s rate added another £0.4bn on the existing loans, and some small adjustments added £0.3bn. Put together, they almost exactly cancel out the bad news.

Waterfall chart showing a 6.4bn pound earnings hit almost cancelled out by paper gains, leaving a 30m pound loss

It’s like your bank balance looking healthy because a refund came in on the same day as a big bill – the balance hasn’t moved, but the bill is still real.

We’ve been here before. In 2021–22 a new earnings model wiped £12.8bn off the value of the loans, and almost nobody noticed, because a change to the Treasury’s rate added £19.8bn at the same time – as we found when we went back through five years of these accounts in March.

There’s one big unanswered question in this year’s figures. Tucked away in a table, DfE says that fixing “bugs” in its earnings model, and updating the information that goes into it, knocked £4.2bn off the value of the loans. The accounts don’t say where that has gone, but the £6.4bn is the obvious candidate.

If so, much of this year’s “graduates will earn less” is really “our model was wrong” – which isn’t much more comforting, given that auditors in Wales and Scotland have already questioned whether the same model overestimates repayments.

Then there’s the model itself. It has some randomness built in – a bit like rolling dice for each graduate’s career – so DfE runs it five times and uses the middle answer. The five runs came out anywhere between £161.8bn and £162.6bn. Running the model again can move the answer by £0.8bn – around 27 times this year’s reported loss.

Chart showing DfE's five model runs ranging from 161.8bn to 162.6bn pounds against the 30m pound reported loss

Piggy in the middle

No real cash changes hands here – DfE is changing its estimate of what it will get back. But there are real implications, because the government still has to count those estimates in its spending, and how it does so matters.

Think of it as two piggy banks. The losses – the day one write-offs, and the knock from lower earnings – go into DfE’s main day-to-day budget, the same pot that pays for schools. The paper gains go into a different piggy bank that the Treasury looks after. So even though the overall loss almost disappeared, the charge to DfE’s day-to-day budget was a record £12.9bn.

Chart showing the cost of student loans to DfE's day-to-day budget, reaching a record 12.9bn pounds in 2025-26

For now, the Treasury tops that up and DfE can’t spend it on anything else, so it doesn’t take money away from schools. But the accounts also explain what happens if the losses get too big. Above a limit agreed with the Treasury, the extra is moved out of DfE’s budget for that year – and DfE then has to pay it back over 30 years, out of the part of its budget that pays for everything else.

It happened in 2020–21 and again in 2021–22. That’s when student loans compete with classrooms.

For the Treasury, the bill is bigger but much harder to see. The government doesn’t have the money it lends to students sitting in a drawer – it borrows it first. Last year that was £16bn once repayments are taken off, all of it added to the national debt.

The interest the government pays on that borrowing is never shown next to student loans anywhere – it just disappears into the overall cost of the national debt. And because the loans are one of the biggest things the government owns, every time their value falls, the chancellor has less room to manoeuvre under the government’s own borrowing rules.

That’s why, as we’ve argued before, the Treasury’s instinct when the loans lose value is to change the rules so that graduates pay back more. Making repayments meaner makes the loans look more valuable immediately. Making them kinder does the opposite.

Off the books

Then there are the mistakes – which may be what held the audit up.

The NAO found that DfE had never recorded the money it had promised students but not yet paid out – mostly the final-term instalments of their loans. It’s a bit like a shop selling gift vouchers and forgetting to write down that it owes the goods.

That came to £9.3bn of promised lending, and DfE should have been recording the loss it expects on that lending – now £3.1bn – for years. The previous year’s figures had to be redone.

DfE also found it had been counting £834m owed by 607,000 people as ordinary student loans, when it should have counted them as overpayments – money paid out to students whose circumstances later changed so they weren’t entitled to it. Having moved that money into the right place, DfE then marked it down by £680m, so it expects to collect only about a third of it.

And the NAO put a formal mark against the accounts over weekend-only courses, where more than 22,000 students received maintenance loans and grants they weren’t entitled to under the rules – around £700m since 2018–19, including £332m of loans and £23m of grants last year alone.

Ministers have decided to write off the grants – £45.5m from 10,596 students – and leave the loans as normal student loans, repaid in the normal way. Separately, the NAO reckons that around £248m of last year’s lending went out without all the proper eligibility checks being evidenced.

For all the talk of write-offs, the amount actually cancelled last year was just £248m – £191m of it because the borrower had died. The big write-offs haven’t started yet. The first Plan 2 balances don’t reach the 30-year mark until the mid-2040s, and Plan 5 loans run for 40 years. Meanwhile, £8.5m went missing because employers took repayments out of graduates’ pay packets and then went bust before passing the money on.

Chart showing student loans cancelled in 2025-26 by reason, led by 191.2m pounds on the death of the borrower

Locked in

Well done if you’ve got this far, which to be honest is more of a congrats to me than it is to you. Down here below the first 2,000 words, you’re at least owed a “so what”.

Well, the Treasury likes to know how much things are going to cost, and likes the idea of being in charge of at least approving, if not crawling all over the justifications for, departmental budgets.

As such, student loans are a problem. Some bright sparks in the last decade removed the ability to control the number of students who go to university. So when more go, ministers can’t take credit. And when even the same number go to the same number of universities and it ends up costing more every year, you get what we’ve had – fee freezes, maintenance that collapsed in value, endless attempts at regulating franchising and so forth.

In the last decade, £9k fees looked like just the latest way of wriggling out of the financial mess the sector was in by transferring the cost to graduates. But because the modelling was wrong – macroeconomic in general, graduate salaries specifically – the belt keeps getting tighter.

This is why student loan terms keep being made meaner, and why DfE ministers are sent out to pretend that fees and maintenance are being indexed.

It’s why the sector’s funding arrangements are so hard to understand, why a government focused on place doesn’t seem to care where universities are, why Plan 2 has gone from the top of Lucy Powell’s in-tray to the wastepaper bin, and why every announcement from DfE looks so wafer thin on ends rather than means.

We’re trapped, exhausted, confused and unloved in a sort of accounting escape room. And until the financial and political bandwidth becomes available to see the system for what it is – farcical – none of the above is going to change.



Source link