The student loan system needs replacing not remixing

This week marks the deadline for written submissions to the Treasury Committee’s student loans inquiry.

We’re expecting press releases from think tanks, lobby groups, and campaign outfits. They won’t all agree. But they’ll likely all, in one way or another, be remixing.

If we think of England’s student loan system as one of those desks in a studio with a bunch of sliders, you can see all the factors folk care about.

One controls how much gets loaned out for tuition, another for living costs, and a fader sets the balance between what the taxpayer puts in and what graduates pay back.

Then there’s how long you’re paying for, whether you pay more when you’re young or when you’re old, whether richer graduates chip in more than poorer ones, and how many loans get issued in the first place.

You can play with how thresholds and balances change over time, whether the government can change the rules after you’ve signed up, and add effects for how fair the whole thing feels to the people inside it.

The Committee itself is only really looking at about half of those – interest rates, threshold indexation, the cost-share between graduates and the state, marginal tax rates, and whether terms should be changeable after the fact.

It’s mostly examining the back end of the system – how loans are priced, described, and repaid – not the front end of how much students borrow or how the state structures support before repayment starts.

Fiddling at the desk

That hasn’t stopped contributors going wider. The Sunday Times, whose “End the Graduate Rip-Off” campaign helped trigger the inquiry in the first place, has sent a submission built around three proposals – end the repayment threshold freeze, reform interest and repayment rates, and introduce a lifetime cap on total interest paid.

The government’s own recent move – a one-year cap on Plan 2 interest at 6 per cent – is even narrower, touching just one slider.

IFS has modelled what happens when you move several at once. HEPI’s Nick Hillman has argued the public debate is over-focused on headline interest and that the real design problem lies in where and how higher rates bite. The Conservatives have proposed scrapping interest above inflation and lifting the threshold.

What actually broke

Government has been fiddling too. The roughly five million graduates who took out Plan 2 loans between 2012 and 2023 ended up on materially worse terms than what Plan 2 was originally supposed to look like.

The threshold at which you start repaying was frozen and then frozen again, dragging more of graduates’ income into repayment each year as wages rise. Maintenance grants were abolished entirely in 2016, pushing living costs onto the loan balance, and interest kept compounding on top.

The combined effect, as both IFS and London Economics now agree, is that average lifetime repayments for 2022 entrants have increased by around £13,400 for men and £16,900 for women compared with the original Plan 2 terms – and the people hit hardest aren’t the highest earners. They’re the low- and middle-earning graduates.

London Economics’ latest modelling says the system now produces a negative RAB charge of around minus 11 per cent for the 2022 cohort – meaning the government expects to get back more than it lent out, once you account for interest. It even calculates a net Exchequer surplus of £679 million on the cohort.

As I noted in January, we’ve stealthily moved from a system that was supposed to be cost-sharing – £4,950 from the graduate, £4,050 from the state – to one where graduates fund everything and then some.

Third time lucky

One new intervention comes from HEPI and NUS, with more modelling by London Economics. Their big idea is “stepped repayments.” Instead of everyone above the threshold paying back at a flat 9 per cent of income, graduates would pay different rates at different earnings levels – 3 per cent on earnings between £12,570 and £27,570, then 5 per cent up to £42,570, then 7 per cent up to £57,570, and then – in an unusual design choice – back down to 3 per cent above that.

The drop at the top is deliberate, designed to stop higher earners from clearing their debt quickly and keeping them in the system for longer so they keep contributing over time. Think of it as a speed bump near the finish line.

If this sounds familiar, that’s because it is. London Economics has proposed versions of stepped repayments at least twice before. In 2023, commissioned by the University of the Arts London, it modelled a scheme with rates of 3, 5, 7, and 9 per cent starting at £25,000, packaged with maintenance grants and sold as cheaper than the post-Augar baseline.

In 2025, commissioned by NUS and UAL together, it produced a scheme with rates of 2, 4, 6, and 8 per cent starting at £12,570, packaged with means-tested maintenance grants of up to £4,224 and a shorter 35-year repayment period, and sold as having “almost identical Exchequer costs” to the current system.

A broken promise

It’s pretty clear that Bridget Phillipson, when in opposition, liked the idea. In June 2023 she wrote that Labour could “reduce the monthly repayments for every single new graduate” without “adding a penny to government borrowing or general taxation.” The promise was specifically about what graduates would see on their payslip each month – lower deductions, no extra cost to the taxpayer.

She told the Labour conference that September that “student finance will be the first to see change” and told a separate audience that the Tory changes were “hammering the next generation of nurses, teachers and social workers” and that a Labour government would “move swiftly to right these wrongs.”

Those promises haven’t been kept. What we’ve got instead is a one-year interest rate cap and a stack of “rapid reviews” that haven’t rapidly reviewed anything. And even the latest HEPI/NUS version of stepped repayments can’t straightforwardly deliver the payslip promise.

This time, repayment starts at £12,570 rather than the current Plan 2 threshold – lower rates, but applied from a much lower starting point. Some graduates would actually see a bigger number on their payslip in the early years, not a smaller one. The promise was about the payslip, and the proposal doesn’t deliver on it.

Square peg, round hole

To keep the idea of stepped repayments while also driving the Treasury’s long-run contribution down to basically zero, the 2026 version has had to harden in several directions at once.

The maintenance grants that earlier versions used to fund support for poorer students have been dropped entirely – presumably the assumption is that the international levy funded version will do for now. The entry point for repayment has fallen from £25,000 in the 2023 version all the way down to the personal allowance of £12,570, and the softer post-study interest rates of 0 to 2 per cent in the 2025 NUS/UAL version have given way to real interest climbing to 3 per cent – harsher than Plan 5, which removes real interest altogether.

So to rescue the aesthetics of stepped repayments without asking the Treasury to put any real money back in, this version is in some respects worse than the original Plan 2 terms, worse than what 2022 entrants are currently on, worse than Plan 5, and worse than both of London Economics’ own previous stepped proposals.

None of those individual changes is necessarily indefensible in isolation. But taken together, they’re the price of fitting the square peg of “nice-sounding graduated rates” into the round hole of “the state pays nothing.”

Pretzels on a crashing plane

There is something seductive about stepped repayments. They feel fairer, look more progressive, and might be easier to sell to an 18-year-old than the current flat 9 per cent above a frozen threshold. If you’re only looking at one slider on the desk – how the repayment rate is structured – it’s a real improvement.

But it’s a bit like being handed a free bag of pretzels while the plane crashes into the side of a mountain.

The system is in crisis because of the interaction between frozen thresholds, compounding real interest, inadequate maintenance, a fee that hasn’t moved in over a decade, and a Treasury that has discovered it can stealthily extract more from graduates whenever it needs to tidy up the public finances.

Rearranging the repayment rate structure doesn’t fix any of that. It just makes the monthly payslip deduction feel a bit less arbitrary while everything else burns.

And anyway, to get there, the new HEPI/NUS stepped-repayment model is designed to produce an Exchequer contribution of approximately zero. The report presents this as a feature – “approximately fiscally neutral,” a split of 0 per cent from the state and 100 per cent from graduates.

Really? I’m old enough to remember when NUS conference used to spend a full afternoon debating whether to campaign for free education – tuition fees abolished, maintenance grants restored to 1979 levels – or for a more pragmatic model of heavily subsidised income-contingent loans.

The politics were fierce. But whatever side of that argument you were on, the underlying assumption was the same – higher education is a public good and the state should contribute to its cost. The question was always how much – never whether.

Now NUS is co-sponsoring a model where the state’s long-run share is zero. It feels like a fundamental concession on the question of who higher education is for – not a tactical compromise.

Picture the scene in a Treasury meeting sometime this autumn. DfE officials walk in with a set of options for the next spending review. The Treasury official across the table says: “Even the NUS thinks the system should be fully funded by graduates. Why are we still subsidising it?”

Bottom of the table

It’s not as if we’re starting from a generous baseline. The OECD’s Education at a glance data has consistently shown the UK at or near the bottom of the table for the public share of higher education funding.

In the most commonly cited figures, only around a quarter of tertiary education spending in the UK comes from public sources – against an OECD average of about two thirds. Norway and Finland are up above 90 per cent. The UK has the highest share of private expenditure on tertiary education in the OECD – most of it coming from students themselves.

Proposing a system in which the public contribution goes to zero would eliminate the last remnants of a public commitment that was already among the thinnest in the developed world.

Worse, it dodges the harder argument about public benefits.

If government wants to influence universities over what to teach, where to teach it, who to let in, how to run their research, and how to serve their local economies – and this government clearly does – then it has to be willing to put money in. It needs a hand on sliders.

The two jars

But the bigger issue isn’t fiddling with one of sliders. When the government lends money through student loans, it doesn’t just hand out cash and hope for the best. Since 2019, at the point each loan is issued, the system splits it into two jars.

Jar one is the money the government expects to get back – this gets recorded as a financial asset. Jar two is the money it expects to write off eventually – and that counts as public spending, right there and then. The idea is sensible enough – if a chunk of every loan is never coming back, the government should own up to that cost immediately rather than pretending it’s all fine and only admitting the truth decades later when the write-offs actually happen.

But the original split between the two jars was based on predictions – forecasts about how much graduates will earn, how the economy will perform, what inflation will do, how people will behave. Those predictions are, inevitably, often wrong. Five years down the line, it turns out earnings growth was weaker than expected, or inflation spiked, or the economy underperformed. The original guess about how much would come back turns out to have been too optimistic – or too pessimistic.

When that happens, the jars get rebalanced. If the updated forecast says graduates will repay less than originally predicted, money moves out of the “asset” jar and into the “spending” jar, and the government has to book that as additional public spending in that year’s accounts – even though no actual cash has changed hands and the write-offs themselves are still decades away.

If the forecast improves, money moves the other way and the government books what looks like a fiscal improvement. The annual deficit can move up or down based on nothing more than someone updating a spreadsheet about what graduates might earn in 2045.

We saw this happen earlier this year, when the Supplementary Estimates revealed that DfE had to book billions in additional write-down charges after adopting a new, more accurate earnings model that showed graduates weren’t going to earn as much as previously assumed.

That’s the involuntary version – forecasts change, jars rebalance, the deficit wobbles. But the same mechanism also applies when the government deliberately changes the rules. And that’s where the incentives get ugly.

Freeze a repayment threshold, and expected repayments go up across the entire existing loan book. Money moves into the asset jar, and because all that writing off you did before now looks like too much, you get a payout in-year too. Raise the threshold, lower interest rates, or do anything that reduces expected repayments, and money moves the other way – the government books an immediate cost.

Because the revaluation applies to the whole outstanding loan book and not just new lending, even modest parameter changes get multiplied across hundreds of billions of pounds of existing debt. A small tweak to a threshold produces an enormous in-year swing in the headline numbers.

As long as the OBR says the numbers stack up, the structural incentive is always there – making things worse for graduates improves the headline fiscal numbers this year, and making things better costs you in the year you do it.

Put bluntly, the accounting system rewards Chancellors handsomely and urgently for gently squeezing graduates, and penalises them hugely and urgently for being modestly generous. That’s not a bug that stepped repayments can fix – it’s a structural feature of income contingent loans of this scale.

It is not a system that can ever be made to work.

Write a new song

Every few years or so, someone re-records “Do They Know It’s Christmas,” and somehow each version is worse than the last. There’s a new production, different artists, and updated arrangement, and another try at that awkward Bono line – but still the same song. And each time, the public likes it a little less. Band-aid maintenance, you could call it.

That’s the danger here. You can fiddle all you like. But if the public are sick of it and it fails to raise the money that’s needed, you need to write a new song. Now.