Meet Mark. He is 47, lives in Arnold, Nottingham. He is a credit risk reporting manager at a credit reference firm on the NG2 business park, earning £63k a year – roughly the top sixth for men his age locally. Mark drives a Skoda Kodiaq on PCP.
Meet Claire, Mark's wife, 44. She is a PA to a director at an energy supplier's city-centre office, on £32k. Claire is a regular reader of Martin Lewis and does the weekly family shop at the big Sainsbury's and tops up with essentials from Lidl to reduce their household spend. Claire drives a 2021 Kia Sportage she owns outright.
They live in a four-bed detached currently worth £450k, bought in 2014 for £300k. They remortgaged in 2021 onto a five-year fix at 2.1%, currently about £891 a month. The fix ends this autumn. About £40k sits in ISAs. Their teenage kids, Jack and Evie, do Gracie Barra BJJ twice a week. Mark's salary-sacrifice pension keeps his adjusted income just a touch under £60k, the point at which the High Income Child Benefit Charge starts, so they keep their full Child Benefit.
They are careful, decent, hard-working people.
At risk
Then, one otherwise unremarkable morning, his firm introduces an AI reporting tool and the MD sends out a restructuring announcement, which is followed by an 'at risk' letter. Then, three months of notice spent job hunting while he is technically still employed. Mark spends his days physically at his desk but endlessly surfing LinkedIn, recruiters give him hope as they need to make sales but after a few weeks of applying there is a visible hiring freeze in all the firms in his area. His blood pressure rises. He looks outside of Nottinghamshire, widening his search criteria on roles – he is 47 and has always had a job.
By contrast, his father Clive worked at the local Calverton Colliery until the mine was closed in 1993, as Mark was 14 and entering his hormonally volatile GCSE years, a formative period in anyone's life – Mark remembers it keenly. The hardship. The times his mother had to take items off the conveyor belt at the cashier's desk at the supermarket, and work extra hours cleaning loos at the city hospital. Mark worked hard at school and did all within his power to ensure it never happened to him. "I have worked so hard to better myself, I am a white collar professional and have transferable skills, I'll be fine" he tells himself as he sends out 20 more applications.
Two weeks pass. Still nothing.
Claire tells him she read in the Daily Mail that, according to the ONS, job vacancies are at their lowest since 2014, outside the pandemic years.
Mark's blood pressure rises further.
Eight weeks
Mark has fixed costs to bear and has dutifully paid into the system all his life never once imagining he would need to rely on the state safety net. He starts asking ChatGPT what he is entitled to – the same technology that took his job now explains his benefits. He winces at the cruel irony. Mark is a composite, before anyone asks. Every number that follows is real, and that is rather the problem.
Mark discovers he is entitled to get Jobseeker's Allowance of £95.55 a week for up to 182 days, based on his National Insurance contributions. £2,484 in total, after 25 years of Class 1 contributions. In a normal year he pays about £15,900 in income tax and employee National Insurance, so his entire unemployment entitlement comes to roughly eight weeks of his own tax, handed back.
Universal Credit is a means test, and Mark's household fails it twice over. Their savings are over £16k, which alone disqualifies any claim. And even with no savings at all, Claire's wages would taper most of an award away while Mark's JSA is deducted from it pound for pound. Whichever door he tries, the answer is the same: come back when the money's all gone.
Handing back the Kodiaq
Once the mortgage resets, their unavoidable costs are about £3,415 a month against about £2,530 of income while JSA lasts, and £2,120 after it ends. The shortfall comes out of their life savings, and it has to because their £40k in ISAs plus about £13k of redundancy pay counts as 'capital', and no household with more than £16k gets Universal Credit. They are, by design, worse off for having done the right thing and put money aside for a rainy day.
Mark used to make £63k, well short of the roughly £77k the Joseph Rowntree Foundation says a family of four needs for a no-frills life (I'd argue the real figure is considerably higher, but that is another story for another day). Claire's salary was what closed the gap, and left a thin margin to save or enjoy life. Take Mark's salary away and, once JSA ends, their income adjusted for family size sits at around half the national median – below the 60%-of-median line the government itself uses to define poverty. Yet the same state that would now count them as poor won't pay them a penny, because they have £53k in the bank.
If they keep the Kodiaq, the family falls below £16k in about two and a half years. If Mark hands it back to the dealership, their costs drop to £2,945 and it takes nearer four years. After that, they receive about £490 a month of UC and then, after a three month wait, a Support for Mortgage Interest (SMI) loan secured on the house. By then Mark will almost certainly have taken a lower-paid job just to stop the bleeding. This is a critical phase change to note, because once a family moves from savings to secured borrowing, getting back is far harder than the arithmetic alone suggests. Mean reversion simply isn't that easy, the assumption that families who fall into debt or bankruptcy can simply 'save their way back to the average' is for the birds – in reality they become economically inert. Run Mark's own numbers: his family's unavoidable costs are around £41k a year. The median full-time salary is £39k. The costs of an ordinary settled life now sit above what the middle of the earnings distribution takes home, and one leaking roof separates a median household from insolvency.
Marc in Lyon
In his Times article Darren Jones, who was Chief Secretary to the Treasury until last September, observed that "the welfare system will not be able to support middle-class workers with good salaries and expensive mortgages if they find themselves out of work." Jones is right, it won't.
If Mark was 'Marc' and lived in Lyon, his unemployment insurance would pay roughly six times more per month, and pay it for three times as long – about £46.6k over 18 months against £2,484 over six. In Portugal, a poorer country, he'd get nearly three times as much a month, for up to three times as long. Of course, those systems are funded by contributions set aside for unemployment which is, nominally at least, what the UK's National 'Insurance' system is meant to do. It mostly funds the state pension he won't see until he is 68 (I suspect, personally, that all those under 50 will never see it). So poor old Mark here pays a continental sized tax bill and gets a safety net thinner than Portugal's.
Many such cases.
Eight weeks of his own tax back in Britain. Nearly three years of it in France.
Former EU powerhouse turned industrial sick-man net-zero suicide note, Germany, and even the relatively poorer Spain would pay more than the UK from day one because Mark has children. In Germany, an unemployed parent gets 67% of their previous net pay instead of 60%. In Spain, the monthly cap rises from €1,225 to €1,575 for a family with two or more children, irrespective of their savings.
In Britain, Mark's JSA is £95.55 a week whether he has two children or none. Child Benefit carries on, but it would anyway. The extra help for Jack and Evie only comes through Universal Credit, which means it only comes once the family's savings drop below £16k, somewhere between two and a half and four years in. In Berlin or Madrid, Jack and Evie would count from day one. In Arnold, they only count once the money has already been spent.
Wide, growing, and nothing for Mark
According to DWP's own figures, 8.4 million people across Great Britain were on Universal Credit in May 2026, the most since it launched in 2013. The TaxPayers' Alliance, working from the same DWP data, puts it at one in five working-age adults in England and Wales, including 1.3 million in their 20s and 2 million in their 30s. Some of the rise is legacy benefits being folded in, but that process is now all but finished, and the new claims keep coming. A third of the people on it are in work, their wages topped up. 4.3 million sit in the 'no work requirements' group – the sick, carers, parents of infants – who the system asks nothing of. That's the system Mark has paid into for 25 years. It's wide, it's growing, and it has almost nothing for him – as one of the people actually paying for it. In a country that places the importance of 'fair play' to levels unseen elsewhere in the developed world, this deeply unfair trade off really smarts.
Punished for saving
The system rewards those who didn't save and punishes those who did. This is the bit that should make every Martin Lewis reader spit out their tea; when UC sizes up Mark's family their 4-bed-detached is invisible and so is his pension pot, so long as he hasn't drawn from it. What the state counts is the cash only; the ISAs, any savings accounts he holds, and his redundancy liquidity – and all that must be expended first, meaning that the family has to burn through the one pot of cash they had built for exactly this rainiest of days. The £40k in ISAs was their emergency fund, the 3-6 months of cover every personal finance columnist tells you to put aside. Meanwhile, Mark has about £290k of equity in his home, ring-fenced. Had he spent the last decade or so overpaying his mortgage, or stuffing money into his pension then the state would help him sooner. I know 'a bit' about this, but if you think you are heading into a crisis, then you sit in cash and wait for opportunities. The UK benefits system treats this liquidity as the first thing to confiscate from you (because it is the easiest) and pays you to lock your money away in an unproductive asset when it rains.
I can already hear readers saying, "Mark should sell his house then".
Ok.
But.
Please also explain to me which lender would give an unemployed man a mortgage on a smaller place?
Walk me through the logic where you downsize and so sell up and rent for a spell, you pay the agent's fees and the removal firm and the solicitor, all for the privilege of paying a private landlord (those who are left) £1500 pcm, whilst Jack and Evie change schools (likely also a downgrade), all in a market where prices are cut and where a £450k property won't sell inside a calendar month. Then concede that the state's own mortgage 'help' is really a loan secured against the property, repaid with interest when it is sold – meaning the equity is only safe until the state wants its cut.
Tempted as I am to opine on the direction of travel for nominal house prices here, it is a much bigger topic, so I will refrain…
Britain used to insure this
So how the hell did we get here? In 1948 we had Beveridge's entirely sensible contributory principle: you paid in, and when bad times came you drew out. In 1966 Labour went further and added an Earnings-Related Supplement, six months of unemployment pay linked to what you used to earn. The minister who introduced it said it would help workers adapt "to meet economic and technological change".
A later government said it existed so people wouldn't be forced to take the first job that turned up. Sound familiar? The Conservatives announced its abolition in 1980, and by January 1982 it was gone. One MP called it "nothing short of robbery" from people who had paid in since 1966. In 1996 JSA halved what was left, from twelve months to six. And from 2013, Universal Credit finished the job: a means test that looks at Mark's ISAs and Claire's salary, and never once at what he paid in for exactly this kind of rainy day. Every party had a hand in it. The last time Britain saw a technology shock coming, it built the insurance Mark needs. Then it took it apart.
Take the first job going
As usual, the system of incentives is all wrong in the UK. The fall from employed to unemployed is steeper in Britain than almost anywhere in western Europe, on the figures above. It strips out the savings a family would need to regroup, and pushes people into the first job available rather than the right one and this has lasting consequences – which we know about because we've tried it. In 1996, JSA halved contributory benefit from twelve months to six. LSE research later found people who lost their jobs just after the change were 2.5 to 3% more likely to end up on Incapacity Benefit and 4 to 5% less likely to be earning a year later, a gap that took four years to close. It costs them, and the Treasury, and therefore taxpayers – for years.
Where AI lands
The Marks of this world are particularly exposed to the way the UK economy functions, and dear reader note this is before so much as a mention of AI. If you look at Goos and Manning's Lousy and Lovely Jobs paper from 2007, they found that UK employment grew at the top and bottom of the pay scale, whilst the middle hollowed out. The hiring trap is that for a £40k role, Mark is overqualified and assumed to be a flight risk by any prudent employer. For a £100k role, he is unproven.
Now let's bring AI back into the frame, its impact on the job market lands squarely in the coordinating middle as a potential future Engels pause for those who, like Mark, compile, reconcile and summarise. For this reason, and by extension, a lot of rules-based white collar jobs are in theory 'at risk' at time of writing. Claire's job is equally 'at risk' so both incomes into this family's household are exposed to the same tech.
Nottinghamshire has seen this before, former coalfields still have 57 employee jobs per 100 working-age residents, against 73 nationally and 88 in the big regional cities (per Sheffield Hallam's State of the Coalfields research). Warehousing now employs over 175,000 people across Britain's former coalfields, close to the pre-strike mining workforce in headcount, if rarely in pay. The replacement jobs did come, decades late, on lower pay, and mostly somewhere else than where they were lost.
Indeed, Britain has been through this before, in the last pause 1790-1840, workers waited around fifty years for wages to catch up with output. South Wales got the worst of it: the Valleys still have 46 employee jobs per 100 working-age residents, and in Maesteg there are 1,500 jobs, no more than a decade ago, with net out-commuting equal to about two thirds of everyone in work. Mark has 21 working years left, his savings last about three of them, and a mortgage that runs until he's 65. Even if AI delivers everything its most excitable advocates promise, on the historical precedent the gains will only reach workers well after their working lives are over. Jack and Evie might just be around for the catch-up. Mark won't be.
What 10,000 Marks cost Nottingham
Thinking back to the article GBTT ran on Britain's internal FX rate, i.e. what a pound in Yorkshire gets you versus London and such, let's game this potential future Engels pause out regionally...
Between his income tax, his NI and the employer NI paid on top of his salary, Mark's job generates about £24.6k a year for the Treasury. Now let's assume there are 10k other 'Marks' across Greater Nottingham and run three scenarios starting with the most optimistic and least likely:
1) Jobs replaced nationally, but not locally – Nottingham loses around £630m a year of gross pay.
2) Rehired one band down at £40k – the tax his job generates falls to about £12,930, a permanent Treasury loss of roughly £11,700 a year each, or, £117m across the 10,000 strong Mark cohort. His take-home pay falls by around £1,100 to £1,200 a month which means no more nights out at Sasha's bar on Front Street. In Robert Allen's account of early industrial Britain, output per worker rose for decades before wages caught up. The economy gets more productive, and Mark gets £1,100 a month poorer.
3) Not replaced at all – HMT loses about £246m a year, JSA costs a one-off £25m never recouped.
Hiring fewer is enough
And the current picture suggests this trend has already started, as a logical response to Government policy and human incentive; vacancies are at 702k – the lowest since 2014 outside Covid pandemic; payrolled employees are down 145k on the year in the August flash estimate (provisional, and the ONS does revise these); youth unemployment is at its highest since 2014 too; London unemployment is 6.8%, the highest of any region and up 0.7 points on the year. Correlation isn't causation and London's problems predate ChatGPT, but the white-collar capital sitting top of the unemployment table is exactly what the front end of this would look like. So AI lands on a labour market that was already stalling in response to higher employer NI and minimum wage costs, and regionally there will be winners and losers (relative terms) as inequity between employment hubs increases. For instance, London is highly exposed per an OECD report that found around 77% of Greater London's employment is in occupations where generative AI could automate – or augment, the OECD is careful to say both – over 50% of all tasks, the highest of any region in its analysis. The UK as a whole was second only to Luxembourg, a country that is little more than a dismal over-regulated, transient, technocratic financial centre with a flag.
The inequity isn't just regional either. In London, per the Mayor's own analysis, women make up nearly 60% of workers in the highest-exposure roles, and 52% of 16-29 year olds are in highly exposed jobs versus 39% of over-50s. Worth noting that 'exposure' doesn't absolutely mean 'job loss', but the OECD does see early signs of declining employment among young, early-career workers in these roles. It also warns that regional differences could widen existing gaps between rural and urban areas. Per the government's own exposure assessment, the most exposed are accountants, finance analysts, credit controllers, payroll managers, HR admin and government administrative roles, with the least exposed in trades – roofers, brickies, gardeners and construction workers. The coalfields actually come out of this quite well, in the same way that an empty house survives a burglary.
Anecdotally, the employers I speak to are not hiring at all and most are not firing either. They are in wait and see mode. The more nimble AI-savvy among them are firing roles in HR, Marketing and rules-based admin to reduce costs and cash burn. The bigger corporates say things like 'we are a family here', and my bet – and it is a bet, the data doesn't show it yet – is that many will need to cut headcount hard to stay profitable as smaller AI-native start ups do the same quality of work for a fraction of the cost and eat into their pricing models. What I can say is that AI was cited in a large chunk of US job layoffs in 2025, and Block (Jack Dorsey's payments firm) cut around half its staff in February this year. Leading indicators, not proof: the Yale Budget Lab and St Louis Fed do not, yet, see displacement at scale. But nobody needs to sack millions at once for Mark's problem to arrive. Hiring fewer replacements is enough. So is removing a layer of administrators, or deciding that five analysts plus AI can do what eight used to do. The statistics will notice that slowly. Mark experiences it in one shocking letter.
So what is the offer?
So where does that leave us? Darren Jones asked "what is our offer to young people if there simply are not enough jobs to go round?" Honestly, I don't see that we have one. Get a vocation? Unless you are an A* student going to a red brick to read history or philosophy or do a STEM, then learning a trade seems far more prudent given the debt implications. I recently met a podcast engineer in his mid twenties on the pre-2023 loan terms, earning £30k with £70k of student debt: this year the loan charges him about £2,870 in interest and he repays £55. He will hand over 9% of everything he earns above £29,385 for thirty years, never clear the balance, and do it in a job that AI audio tools are already learning to do. No one rational takes that deal.
There is another version of the question of course, which is what our offer is to someone like Mark.
We have a whopping 8.4 million people already claiming UC per DWP data, a caseload equivalent to just under a quarter of the UK workforce. Only 1.6m of those 8.4m people are looking for work; 4.3m are not expected to work. Meanwhile, Mark in Arnold gets £2,484 and Marc in Lyon gets about £46.6k under the same conditions; eight weeks of his own tax back in the UK, nearly three years of it in France, paid over 18 months.
One fix to this could be an automation levy that funds some form of wage insurance. The levy is the Layoff Trap paper's idea, spending it on wage insurance is mine. You basically would get a temporary top-up as a % of your salary (perhaps capped over a certain figure) that can at least afford you some breathing room when a displaced worker needs to take a lower paid job, it keeps people working. If Claire loses her job too (and run this theory to its natural conclusion and she absolutely does), then this family qualifies for state aid sooner and there is less to protect. Their teenage kids, Jack and Evie will never be able to afford a house on the street they grew up on, that is, until the seismic downward repricing in the housing market starts to bite.
The grim reality is that Britain is now a low-wage, high-cost, asset-inflated, transfer economy optimised for the bottom quintile and subsidised by the middle and top of the income distribution. Ultimately, we have created a system where the only way to survive is to be destitute enough to qualify for aid, or rich enough to ignore the cost. The incentives are 180 degrees wrong, with everyone in the middle, like Mark, being cannibalised.
Somewhere around year three, Claire is standing at the Sainsbury's checkout and the total is higher than she expected. She takes a few things off the belt and puts them quietly to one side. The way Mark's mum did in 1993. He spent his whole adult life making sure it would never come to this.
Notes & Sources
- GOV.UK, Jobseeker's Allowance: what you'll get: new style JSA of £95.55 a week for claimants aged 25 and over, for up to 182 days.
- GOV.UK, Universal Credit: money, savings and investments: the £16,000 capital limit and the disregard of the home you live in.
- GOV.UK, Support for Mortgage Interest: SMI as a loan secured on the home, and the waiting period for Universal Credit claimants.
- Joseph Rowntree Foundation, A Minimum Income Standard for the United Kingdom in 2026: the income a household needs for a minimum acceptable standard of living.
- DWP, How low income is measured: the 60%-of-median relative poverty line.
- ONS, Annual Survey of Hours and Earnings 2025: median full-time annual earnings.
- Report of Darren Jones's article in The Times, September 2026: the former minister's warnings on AI, jobs and the welfare system.
- Unédic, Allocation d'aide au retour à l'emploi (ARE): French unemployment insurance rules and duration.
- Bundesagentur für Arbeit, Arbeitslosengeld: German unemployment benefit at 67% of net pay for parents, 60% otherwise.
- Seguridad Social, prestación contributiva por desempleo: Spanish benefit caps by number of children.
- DWP, Universal Credit statistics to May 2026: 8.4 million people on Universal Credit, conditionality groups and in-work share.
- TaxPayers' Alliance, 1 in 5 working-age Brits in England and Wales receiving Universal Credit: Universal Credit claimants by age band.
- Hansard, Reduction and Abolition of Earnings-Related Supplement, 21 May 1980: the Commons debate on abolishing the Earnings-Related Supplement.
- House of Commons Library, The Jobseekers Bill (Research Paper 96/5): the 1996 cut in contributory entitlement from twelve months to six.
- Petrongolo, B. (2009), The long-term effects of job search requirements: evidence from the UK JSA reform, Journal of Public Economics: effects of the 1996 JSA reform on incapacity benefit take-up and earnings.
- Goos, M. and Manning, A. (2007), Lousy and Lovely Jobs: The Rising Polarization of Work in Britain, Review of Economics and Statistics: the hollowing out of middle-paid employment.
- Allen, R. C. (2009), Engels' pause, Explorations in Economic History: output per worker and real wages in early industrial Britain.
- Sheffield Hallam University, The State of the Coalfields 2024: employee jobs per 100 working-age residents and warehousing employment in former coalfields.
- ONS, Labour market overview, UK: September 2026: vacancies, payrolled employees and youth unemployment.
- ONS, Regional labour market: September 2026: London unemployment rate.
- OECD, Job Creation and Local Economic Development 2024: The Geography of Generative AI: regional exposure of employment to generative AI.
- Greater London Authority, London's workforce exposure to generative AI: exposure by sex and age in London.
- Department for Education, The impact of AI on UK jobs and training (2023): occupational exposure to AI.
- Hemenway Falk, B. and Tsoukalas, G. (2026), The AI Layoff Trap: the case for an automation levy.
- Straight Arrow News, Artificial intelligence tied to more than 50,000 layoffs in 2025: Challenger, Gray & Christmas counts of US job cuts attributed to AI.
- CNN, Block lays off nearly half its staff, February 2026.
- Yale Budget Lab, Tracking the impact of AI on the labor market.
- Federal Reserve Bank of St. Louis, Six questions about employment and generative AI.
- House of Commons Library, student loan interest rates and repayment thresholds: Plan 2 repayment threshold and interest.