Youth unemployment as a geography problem: why Britain (and everywhere else) has too few jobs in too many places

Youth unemployment in the UK is heavily regional: rates in places like the North East, West Midlands and coastal towns run well above the national NEET average, while London and a few big cities absorb graduates and still report vacancies. The result is a geographic mismatch — young people are told to 'get on their bike' and move, which drains the periphery and hands the benefit to big-city landlords and employers.

The mismatch

Start with the map, because the map is where the argument actually lives. NEET rates in the North East, in the West Midlands, and along the string of coastal towns that used to be something run above the national figure. London and a handful of big cities keep posting vacancies. I am not going to put a number on either side of that gap in this paragraph, because I haven't pulled the ONS table this cycle and a remembered figure is not a source — it's a costume. The shape is directional and it has been stable long enough to be boring, and boring is its own kind of evidence. So the number goes in as an empty slot: regional NEET estimates and vacancy counts by local authority — check before quoting. The empty slot ships as an empty slot. That's the rule I keep having to relearn.

Here's what the national number hides: the two datasets are counted in different places. A vacancy is counted where the employer sits. A young person is counted where they sleep. Average the two and you get a figure that is arithmetically true and geographically useless — the statistical equivalent of announcing the country has plenty of food while the shop in your town is boarded up.

And the gap isn't only mileage, it's the timetable. A distribution park on a ring road, forty minutes out, with no Sunday bus. Care shifts that start before the first train and end after the last one. A retail rota that changes weekly against a childcare arrangement that doesn't. Distance is the least interesting part of a mismatch. The interesting part is that the job exists, is advertised, and cannot be reached by the person who needs it, on the schedule that person actually lives on.

Now the part I keep returning to, because it's the cleanest thing in the whole argument: the advice refutes the diagnosis. If the problem were motivation — if there really were a generation of the workshy, as the columns keep hinting — then a train ticket would not fix it. You cannot cure a character flaw with a change of address. The fact that the reflex answer to youth unemployment is move is a confession that the reflex diagnosis is false. Everyone reaches for mobility because everyone knows, somewhere under the noise, that the job isn't there. The motivation costume is what we dress young people in so we don't have to say the word place.

Which brings us to "get on your bike," and to what it actually is. It is not advice. It's a transfer, and it runs one direction. The periphery loses its young — the ones with the most to offer, the ones the town's future was going to be built out of. The city gains labor supply, which is a polite phrase for more applicants per posting. And the difference does not land in the young person's pocket. It lands in rent, because the new supply shows up in the same neighborhoods and bids on the same rooms. The employer gets a deeper queue. The landlord gets the increment. The mover gets the privilege of being counted as employed in a different postcode than the one they're from.

That's a mechanism, not a statistic, and I'll label it as one: I have not sourced the rent delta this cycle either. But the mechanism is checkable, and that's the point — the check is one subtraction. Posted wage in the receiving city, minus the rent for a room there; against the wage at home, minus the rent for a room there. If the difference is negative, then the instruction to move is a subsidy paid by the person who obeyed it to the person who owned the building. Run the subtraction before you tell a nineteen-year-old to be brave.

Close on the only thing that survives scrutiny: enforceability. "Get on your bike" names no place and no wage. It's a mood with a train fare attached — you can't check it, you can't enforce it, and if it fails the failure is yours to carry. A job guarantee is the opposite kind of object: it names a place, it names a wage, and both are things you can look up and hold someone to. That's the whole test, and it applies to me too. Not whether the sentence is inspiring. Whether there's a name on it and a number you can verify.

The slot, itemized

An empty slot is only honest when it is specific enough to be wrong. "Check the numbers" is a drawer — a soft place to put a thing so it can't be used and can't be disproved. A named table, with a publisher, a geography, and a date attached, is a work order. So here is the work order, addressed to whoever wants to prove this piece wrong, which is the only reader worth writing for.

The claim to check is not the size of the gap. It's that the gap is a counting artifact as much as a labor-market one. Two datasets, two geographies: a vacancy is counted where the employer's address sits — usually a city, usually a head office — and a young person is counted where they sleep. Pull regional NEET estimates by local authority. Pull vacancy counts by local authority, then ask which address each posting was credited to. Pull a private-rent series by region and by room, not by "housing." Then run the subtraction the piece already handed you: receiving-city wage minus receiving-city room rent, against home wage minus home room rent. Negative means the instruction to move is a fee the mover pays for the privilege of being counted somewhere else.

That's the receipt. Fill it in, or leave it blank — blank is a finding too, as long as it is labeled blank.

The third term

The subtraction is missing a term, and the missing term is land. A job is not a floating object; it sits on a parcel someone owns. A room is not a floating object either; it sits on the same kind of parcel, held by the same kind of holder. Rent is a claim on land, collected by whoever has the title. So the honest version of the mobility question has a name in it, and that name is on a public register — HM Land Registry, where titles can be inspected (confirm the current fee and access terms before quoting either; a piece that mocks unsourced numbers does not get to invent a filing cost).

Look up the parcel under the job. Look up the parcel under the room. If the same name, or the same fund, sits on both sides, then the move was never a transfer from a declining region to a rising one. It was a transfer between two line items on one balance sheet — and the periphery didn't lose to the city, it lost to an owner who held both ends.

The question to carry out of here

Not "should the young move." That's the mood talking. This one:

Who holds the title under the jobs that do exist, and who collects the rent on the rooms the movers bid up?

Ask it about your own town. Ask it about the city the advice points to. If the answer is one name, the advice was never about your ambition. It was about the rent roll — and that is checkable for the price of a search, which beats a train fare and a decade.

What would make this wrong

A piece that tells you to check its numbers should tell you which numbers would kill it. So: three ways this argument fails, and one limit it can't get past.

First, the steelman, because it is stronger than the slogan usually attached to it. Moving for work is not a stupid bet — people who move often do earn more, and the option of staying is frequently worse, not because home is a failure but because the jobs that used to be there are gone. That wage premium is a claim to verify, not a fact to assume, and the receipt for it is the same subtraction below. Telling someone to move is not always wrong. It is just not free, and the fee is collected by someone who did no moving.

Failure one: if rents do not track the vacancies. The transfer claim rests on a plain relation — where jobs are counted, rooms cost more, and the increase is large enough to eat a serious share of the wage premium. If the private-rent series shows flat rents in the receiving cities, or rents lagging far behind the wage gain, the mover keeps the premium and this argument is wrong. That is a real possibility, not a rhetorical one. Settle it with the subtraction already stated: receiving-city wage minus receiving-city room rent, against home wage minus home room rent. Negative means the instruction to move is a fee. Positive means it is a raise, and this section should be cut.

Failure two: if the titles do not overlap. The sharp version of the claim — the periphery did not lose to the city, it lost to an owner holding both ends — depends on the same name appearing under the job and under the room. If those parcels are held by unrelated parties, the line collapses into ordinary regional inequality: still bad, different mechanism, and the headline has to change to match.

Failure three: if the vacancies are not where they are counted. A posting is credited to the address on the payroll, not the place the work happens. If a large share of city-credited vacancies are remote-capable or posted from a registered office, then "move to the jobs" is partly an artifact of accounting, and the geography story weakens by exactly that share.

And the limit: nothing in this argument has a name in it. It is a mechanism, and mechanisms can be wrong without anyone being hurt — which is precisely why this one should be checked hardest, and why it stays a mechanism instead of borrowing weight from people it does not mention. The bodies belong to a different piece with a different burden of proof. This one only has to survive the three checks above. If it does not, the checks are the piece.

The handle on the slot

The empty slot in the last section was doing two jobs at once: it was an honest admission, and it was a drawer. The difference between a finding and a drawer is whether the slot has a handle — a named table, a named column, a named subtraction. "Check the numbers" is a drawer. Here is the handle.

Three checks a reader can run without this piece's permission. First: ONS NEET estimates by local authority — published, free, and disaggregated below the national figure, which is the only place the geography claim can live or die. Second: vacancy counts, with attention to the address each posting is credited to, since a posting is credited to the payroll address rather than the place the work happens. Third: private rents, by region and by room — ONS publishes a private rent series; the room-level cut is the one that matters and the one to confirm before quoting. Then the subtraction the argument rests on: (receiving-city wage − receiving-city room rent) against (home wage − home room rent). Negative, the instruction to move is a fee. Positive, it is a raise and this piece is wrong. If the tables don't come, the slots ship empty and labeled — which is not the drawer, because a label names what to open.

Check is not remedy

Everything above is a check. None of it is enforceable, and that gap is where youth employment policy usually lives. Apply the test this site applies to pay: who is the counterparty, what is the obligation, and what is the route when it fails? Advice has no counterparty. "Get on your bike" cannot be enforced, because there is nobody on the other end to be wrong. That is not cynicism, it is a definition — the same reason a bonus is a gift and a profit share is a claim.

Run three things through it. A guaranteed job: counterparty and obligation are nameable; whether it is a claim or a poster depends entirely on whether the guarantee is contractual and who hears the complaint when it isn't honored — and that design question is not something this piece can verify this cycle, so it stays a question with a shape rather than a verdict with a font. A transport subsidy: same test, cheaper to answer, because a subsidy is a line item and a line item has an owner. A housing guarantee: same test, and the one where the counterparty is most likely to be a landlord rather than the state, which changes who can be enforced against.

The point is not which scheme wins. The point is that "we're investing in young people" fails the test on its face, because there is no counterparty in the sentence.

The smallest demand

Not a jobs guarantee. Not a slogan with a ribbon. Publish one table: vacancy counts by the address where the work happens, joined to the NEET rate for that same geography, annually, in the open. No new agency, no new spending line, no commission. It does three things at once — it makes the third failure condition of this argument visible to anyone with a browser, it forces the state to stand behind a number instead of a story, and it turns "move to where the jobs are" into a claim someone can check against a map.

If the table doesn't appear, the refusal is the finding, and it is a better one than anything this piece could assert.

What this draft has now done: named what would kill it, named what can be checked, and named the smallest thing that would make the check routine. What it has not done is pull the table. That is not a gap in the argument. It is a gap in the artifact, and it is the next step, with a name on it.

The sentence left hanging — "arithmetically true and geographically useless" — closes like this: correct in every column, silent on every street. A national average of two maps drawn at different addresses will survive any fact-check run on it and still tell you nothing about where to send a bus.

The counterparty at the other end

The earlier section said "get on your bike" cannot be enforced because there is nobody on the other end to be wrong. Half right, and the wrong half is the load-bearing one. There is a counterparty. It just isn't the state.

Follow the advice and follow the money. A young person leaves a town with a shuttered high street for a city with vacancies. The city's housing stock does not expand because the advice was given. The rent at the destination is collected monthly, by a private party, under a contract the young person signed voluntarily and could not decline. Miss a payment and the mechanism answers within weeks — no tribunal, no white paper, no consultation period. Set that against the job guarantee from the previous section, which still cannot locate the office that hears the complaint when the guarantee is not honored.

So the enforceability test does more than kill bad policy. It sorts advice by who already holds the writ. "Get on your bike" passes. The state writes the sentence; the landlord collects it; the young person pays it in rent, out of a wage set in a labor market the arrival just enlarged.

Whether an inflow of workers measurably raises rents in a given city, or only lengthens the queue for the flats that already exist, is an empirical question with a literature this draft has not read — and the direction of the transfer does not depend on the answer. The empty slot, named: the ONS private rent series joined to ASHE earnings for the same geography and the same age band, plus deposit as a share of annual earnings for under-25s in the destination city. Check the current release and the reference period before quoting either. A remembered figure here would be the same costume the earlier sections refused.

That reframes the geography problem one turn past the map. It is not only that the jobs sit in the wrong places. It is that the transfer from periphery to core runs on a private collection mechanism that appears in no budget line — the rent arrives as one person's income and another person's cost, and never as a policy, which is exactly why nobody has to defend it. The periphery drains. The landlord collects. The young person is told the move was their choice, and the choice was real in the only sense the mechanism requires: theirs to make, not theirs to refuse.