Is student work on the increase or not?

He said that despite what people think, fewer university students are working these days, not more. He pointed to official government statistics to back this up.

This isn’t just about those in school/college. Despite what you hear about more uni students working these days the opposite is true – 18-21 year olds studying for degrees have seen chunky falls in employment. So have those studying for other qualifications.

[image or embed]

— Torsten Bell (@torstenbell.bsky.social) 4 January 2026 at 10:49

Bell is nominally a junior minister, but his real role in government goes well beyond pensions. Since August 2025, Rachel Reeves has given him additional responsibility for economic policy and put him in charge of leading budget preparations.

The New Statesman reported it as a significant expansion of his role, with Treasury sources stopping just short of calling him “deputy chancellor.” The bookmakers have him second favourite to be the next chancellor of the exchequer.

When Bell makes confident public claims about labour market data, it isn’t backbench musing – it’s a signal of how the Treasury’s most influential policy mind is reading the evidence.

But there’s a big problem with his claim. The statistics he’s relying on were never designed to count working students – and when you look under the bonnet, they’re really bad at it.

Worse, three separate surveys that were actually designed to measure student employment all say the opposite of what Bell claims.

Two surveys, two stories

There are two main sources of information about whether students work during term time.

The first is the Labour Force Survey, run by the Office for National Statistics. This is a massive government survey that tries to measure how many people across the whole country are in work, looking for work, or not working. It covers everyone – builders, bankers, teachers, retirees, teenagers – and students are just one small slice of the people it talks to.

The second is HEPI’s Student Academic Experience Survey, which is specifically designed to ask university students about their lives. It talks to thousands of students directly and asks them, among other things, whether they do paid work during term time.

These two surveys give wildly different answers. The Labour Force Survey suggests student employment has been broadly flat or falling. The Student Academic Experience Survey says it has roughly doubled – from about a third of students working in 2012 to about two-thirds now.

Bell picked the first one. But there are very good reasons to think the second is closer to the truth – and it isn’t the only purpose-built survey telling that story.

The government’s own data

The Student Income and Expenditure Survey is a large, detailed study commissioned by the Department for Education itself. It’s run by NatCen Social Research and the Institute for Employment Studies – serious research organisations – and it samples students directly from university and college records.

It has been running at roughly three-year intervals since the mid-1980s, although ominous, there’s no current sign of an update to the last one – which was carried out in the dying months of the pandemic.

That most recent edition, covering the 2021/22 academic year, found that the proportion of full-time students in paid work alongside their studies increased from 52 per cent in 2014/15 to 58 per cent in 2021/22. Earnings from paid work contributed a larger share of total student income – up from 17 per cent to 24 per cent. Median earnings for those in work rose by 27 per cent in real terms. Average hours worked also went up.

An earlier wave, covering the academic year ending 2012, found that 70 per cent of students undertook some form of paid work during either the academic year or the previous summer vacation.

So the government’s own dedicated student finance survey shows student employment rising. That makes three purpose-built instruments – SIES, SAES, and ONS’s own administrative data research – all pointing in the same direction.

Against that, Bell is relying on one subgroup estimate from a general household survey that was never designed to measure student working.

Wrong net, wrong fish

Imagine you wanted to know how many fish were in a lake. You could drag a huge net across the whole lake and count everything you caught. That’s basically what the Labour Force Survey does – it’s a giant net designed to catch the whole working population.

But what if your net had holes in it that were exactly the size of one particular type of fish? You’d keep catching everything else, but that one species would slip through. Your count of the total fish population might still be roughly right, but your count of that specific species would be terrible.

That’s what happens with students in the Labour Force Survey.

The survey works by picking household addresses from the Royal Mail’s list of addresses. Interviewers knock on doors – or phone people up – and ask everyone who lives there about their work. If you live in an ordinary house or flat, you’re on the list and you might get picked.

But students often don’t live in ordinary houses or flats. Many live in halls of residence – big purpose-built blocks run by universities or private companies. Those buildings aren’t on the normal address list.

The survey tries to get around this by counting hall students through their parents’ homes instead. If your mum and dad’s house gets picked for the survey, and you’re away at university in halls, they’re supposed to mention you.

You can probably already see the problems with this.

Invisible internationals

If you’re an international student, your parents don’t have a UK address. They’ll never be picked for the survey. And your hall isn’t on the address list either. So you’re essentially invisible.

The reason this matters is that international students are actually more likely to work than home students – 77 per cent compared to 65 per cent. If the survey is missing a big chunk of working international students, it will undercount working students overall.

The halls gap

Over the last decade, there has been a massive expansion in privately run student accommodation – the kind built and managed by companies rather than universities. Between 2014 and 2024, the number of students in private halls grew from about 118,000 to 206,000.

ONS has admitted that the workaround it uses to count hall students hasn’t been updated since it was first created. The list it relies on reflects what student accommodation looked like years ago, not what it looks like now. So as private halls have mushroomed, the survey hasn’t kept up.

Just 457 people

Perhaps the most damning issue of all. When I dug into the actual raw data behind the Labour Force Survey, I found that in a full three-month period in 2023, the total number of full-time first-degree students aged 18–24 who were actually recorded was… 457 people.

That’s the real number – not thousands, not tens of thousands, but 457 actual human beings, whose answers get multiplied up to represent about 2 million real students.

Each person in the sample represents roughly 5,000 real students. If just 30 of those 457 people happen to switch from “I have a job” to “I don’t have a job” between one quarter and the next, the official employment rate swings by six or seven percentage points.

What looks like a “chunky fall” in student employment could easily be a few dozen people having a quiet week.

Not who you’d expect

Those 457 aren’t a representative slice of students, either. When you break down where they actually live, about 57 per cent are at their parents’ home.

In real life, only about a third of students live at home. So the survey has far too many “living with mum and dad” students and far too few “living in halls” or “living in HMOs” students.

That isn’t random. It’s built into the design. The survey is good at finding people who live in family homes because that’s what it was made to do. It’s bad at finding people in rented accommodation because that was never really its job.

Wrong question

The Labour Force Survey asks a very specific question – did you do at least one hour of paid work in the last week? If you did, you count as employed. If you didn’t – even if you normally work, but happened to have no shifts that particular week – you count as not employed.

The Student Academic Experience Survey asks a different question – do you do paid work during term time? The Student Income and Expenditure Survey asks about paid work alongside studies across the academic year.

If you work some weeks but not others – which is very common for students with casual, zero-hours, or gig-economy jobs – you’d say yes to both of those questions. But you might say no to the Labour Force Survey if your interview happened to fall in a week when you had no shifts.

This is significant because the growth in student working appears to be concentrated in exactly this kind of irregular, low-hours work. More students are picking up shifts here and there, doing a few hours on a delivery app, or working occasional weekends. The surveys that ask students directly pick this up. The Labour Force Survey misses much of it because it only asks about one specific week.

Ask mum

Even when the parental-household workaround does work – even when a student in halls is successfully picked up through their parents’ address – it’s often the parent who answers the survey questions, not the student. Government data shows that parents answering on behalf of their adult children is very common for this age group.

The survey needs to know whether the student did at least one hour of paid work in one specific week. If your son or daughter is living two hours away and picking up the odd shift through an app, doing a few hours behind a bar, or getting irregular work through a zero-hours contract, would you know? Would you know exactly which weeks they worked and which they didn’t?

Parents probably know if their child has a regular, steady part-time job. They’re much less likely to know about the kind of casual, irregular, low-hours work that appears to be driving the increase in student employment. If a parent doesn’t know, they’ll either say “no” or give their best guess – and in either case, the survey will undercount students who work.

This isn’t a small issue. It compounds every other problem. The survey already struggles to find students in halls. When it does find them, it’s often asking the wrong person. And the wrong person is least likely to know about exactly the type of work that has grown fastest.

The Milburn connection

Bell’s thread wasn’t made in isolation. It was referencing the government’s Milburn Review – a major independent investigation led by former Health Secretary Alan Milburn into why nearly one million 16 to 24 year olds aren’t in education, employment or training. The review’s interim findings are due in spring 2026, with a final report in the summer.

The framing of Bell’s argument matters because it feeds a particular narrative about youth inactivity – one in which students are supposedly less engaged with the labour market and the picture is getting worse across the board. But the same Labour Force Survey data that Bell is drawing on tells a more complicated story when you look at the structure carefully.

In the early 1990s, about 17 per cent of 18 to 24 year olds were classified as full-time students. Now it’s about 33 per cent. That’s a massive compositional shift. When you move a large chunk of more academically engaged young people into the “student” category, you change the make-up of whoever is left.

Think of it this way. If you have a class of 30 children and you send the ten strongest swimmers to a swimming squad, the average swimming ability of the remaining 20 will go down – not because anyone has got worse at swimming, but because you’ve removed the strongest swimmers from the group.

Something similar happens with the non-student 18 to 24 population. As higher education has expanded, the young people who remain outside full-time education are increasingly concentrated among those who left education earlier, those with lower prior attainment, and those with health conditions or other barriers. The non-student group has become smaller and more selected.

Even if nothing had changed in how individuals behave, the average inactivity rate of non-students would tend to rise simply because its composition has shifted.

The data bear this out. Non-student inactivity among 18 to 24 year olds has risen from 13.4 per cent to 17.3 per cent since the early 1990s. That’s a real change and it matters. But it’s a much smaller movement than the headline figure suggests.

Overall inactivity for 18 to 24 year olds has gone from 21.8 per cent to 29.9 per cent – but the bulk of that increase is explained by the fact that a much larger fraction of young people are now in full-time education, and students are more likely to be economically inactive by definition.

The displacement squeeze

What connects the student employment story directly to the youth inactivity crisis is this.

The HEPI Student Working Lives report, published in November 2025 by the University of Lancashire in partnership with HEPI, the University of Liverpool, London South Bank University and Buckinghamshire New University, put it bluntly. It challenges the “myth” of the full-time student.

Two-thirds of students work to cover their basic living costs. A quarter work to support their families. Students in the survey were working an average of 18 hours a week alongside their studies – with 29 per cent working more than 20 hours – producing what the report describes as a 50-hour week that leaves little room for anything else. Paid work, the authors conclude, is now a necessity rather than a choice for the majority of students.

The report also tells us where students are working – mostly in health and social care, retail and hospitality, with regional labour markets shaping the mix. In Lancashire, where health and social care accounts for 15 per cent of all jobs, a higher proportion of students work in care. In Liverpool, where hospitality accounts for around 10 per cent of city-region employment, more students work in pubs, restaurants and hotels. These aren’t niche findings.

They describe the sectors that dominate the entry-level labour market for all young people.

Collapsing sectors

Those sectors are now in structural collapse. The Centre for Retail Research recorded over 200,000 retail job losses and 17,000 store closures in 2025 – following 170,000 losses in 2024. In hospitality, over 3,300 businesses went insolvent in 2025 and the sector is now 14 per cent smaller than it was before the pandemic, with two licensed venues closing every day.

Nearly 9,000 hospitality jobs were lost in December 2025 alone – a month when the sector would normally be hiring for the Christmas rush. UKHospitality estimates that job losses could reach 100,000 by the time of the next budget, driven by employer national insurance rises, business rates increases and the ongoing cost squeeze.

As retail and hospitality jobs disappear, students – who are mobile, motivated and willing to work irregular hours – appear to be adapting by moving further into sectors like social care, warehousing, delivery and manufacturing that still have vacancies and offer the shift-based flexibility that fits around inflexible lectures.

The Student Working Lives data already shows health and social care as a major employer of students, and care work has chronic staff shortages that actively recruit from the same pool of young workers.

But these are also precisely the sectors where non-student young people might have otherwise found work – and precisely the sectors that government programmes designed to reduce economic inactivity tend to target.

180,000 invisible workers

On top of the displacement effect, there’s the sheer volume. Between 2018/19 and 2023/24, international student enrolments in UK higher education grew by nearly 48 per cent – from about 496,000 to over 732,000, an increase of 236,000 people.

If international students work at the rate the SAES suggests – 77 per cent, compared to 65 per cent for home students – that’s potentially around 180,000 additional working students who entered the youth entry-level labour market over five years.

They need work for exactly the same reasons as home students – to cover living costs that maintenance support doesn’t reach – and they face additional constraints that channel them into precisely the kind of flexible, shift-based, low-status work that dominates the entry-level market – visa hour limits, employers who undervalue overseas experience, and what the Student Working Lives report describes as a perception that their prior skills and social capital don’t count.

The critical point is that the Labour Force Survey can’t see any of these workers. International students whose parents live overseas are structurally excluded from the LFS sampling frame – they aren’t in a UK parental household, and their hall or private rented flat may not be on the main address list. And they’re not weighted up to match HESA – weighting up in the LFS is done to match the Census!

So around 180,000 additional working young people have entered the labour market over five years, competing with non-student young people for the same entry-level jobs – and the government’s primary labour market survey doesn’t know they exist.

A vicious circle

The circularity at the heart of this is bleak. Maintenance support has failed to keep pace with the cost of being a student. HEPI’s Minimum Income Standard for Students, published in 2025, found that even students receiving the highest levels of maintenance support must work over 20 hours a week at the national minimum wage just to meet a basic acceptable standard of living.

The full-time student, as a category, has effectively ceased to exist for most people. Students must work. And because they must work, they enter the same entry-level labour market as non-student young people – competing for the same shifts in the same care homes, the same warehouse packing roles, the same delivery slots.

If the real level of student employment is much higher than the government thinks – and all the evidence says it is – then students are absorbing a much larger share of the youth entry-level labour market than the official data implies.

That displacement effect makes it harder for non-students to find work, pushing up measured NEET rates. And it isn’t some mysterious social pathology driving this. It’s the direct consequence of a funding model that doesn’t give students enough money to live on.

A problem of its own making

So the government faces a problem it has partly created. By allowing maintenance support to erode in real terms, it has forced students into the labour market at unprecedented scale. Those students are now competing with non-student young people for a shrinking pool of entry-level jobs in collapsing sectors.

The compositional effect of higher education expansion has made the non-student group smaller and more selected. The displacement effect of mass student employment – compounded by a surge of 236,000 additional international students over five years, almost none of whom the LFS can see working – has made it harder for that group to find work.

And the sectoral collapse of retail and hospitality has squeezed the available jobs even further, pushing everyone into a narrower range of alternatives where competition is more intense. All three effects push non-student inactivity upward – and none of them show up properly in the data the government is using.

Three separate surveys designed to ask students about their working lives all tell the same story. The Labour Force Survey tells a different story, but that story is built on an unweighted sample of a few hundred students, drawn disproportionately from parental homes, missing large numbers of hall residents and international students, using a question that doesn’t capture irregular work, relying heavily on parents who may not know the answer, and weighted up to represent millions of people using a method that doesn’t correct for any of these problems.

When Bell says “despite what you hear about more uni students working these days, the opposite is true”, he’s choosing to trust the butterfly net over the fishing rod – the giant, general-purpose survey that was never designed to count student employment over multiple smaller, purpose-built surveys that were.

The detailed evidence from the Labour Force Survey’s own microdata suggests that its student employment estimate is too fragile, too skewed, and too blunt an instrument to support the confident claim he’s making.

And the government’s own dedicated student finance survey – commissioned by a fellow government department – contradicts him directly.

That matters more than it might for most junior ministers. Bell isn’t just the pensions minister – he’s the person Reeves has put in charge of economic policy inside the Treasury, the minister the bookmakers think will be the next chancellor.

If this is how the government’s most influential economic mind is reading the data on young people and work, there’s a real risk that the Milburn Review and the policy that flows from it will be built on a misdiagnosis, let alone policy on maintenance for students.

The story isn’t that students have stopped working. The story is that more students than ever are working – many of them 18 hours a week or more, in care homes and warehouses and on delivery apps, just to cover their basic living costs – and that by doing so they’re competing directly with the very group of non-student young people the government is trying to help.

Getting the student employment numbers wrong doesn’t just produce a bad chart. It obscures a mechanism that is actively making the youth inactivity crisis worse.