The State Is Eating the Economy

Author: GBTT Research

Date: March 2026

---

1. NIESR Warned Us. Nobody Listened.

In Spring 2023, Paul Mortimer-Lee of the National Institute of Economic and Social Research published a short, uncomfortable paper: "Is the UK public sector too large?" The answer implied by his data was yes — and by a considerable margin.

Three years on, every narrow measure of public sector size has deteriorated further. The state has added headcount at every level. Manufacturing has continued to shed jobs. Unemployment has risen. Youth unemployment is at its worst in a decade. The crowding-out dynamic Mortimer-Lee documented has not reversed. It has accelerated.

This article takes the NIESR baseline and updates it with the latest available data. The picture is not ambiguous.

---

2. The NIESR Finding, Plainly Stated

Mortimer-Lee used a broad employment measure — counting public administration, education, and health and social work as the public sector, on the grounds that it is not who employs labour that matters for resource allocation, but what that labour produces. On that definition, the public sector's share of all jobs rose from 25% in 1997 to 33% in 2022. That is one in three workers, up from one in four.

The scale of the absorption is the striking part. Over the same period, total employee numbers rose by 6.5 million. Public-sector-dominated industries accounted for 4 million of those jobs — 60% of all employment growth over 25 years. The private productive economy, including manufacturing, received the remaining 40%.

The consequences for manufacturing were direct and measurable. In 1997, manufacturing employed just over 4 million people — three times as many as were in public administration, and more than double those in education. By 2022, manufacturing and public administration were at rough parity in headcount, and manufacturing's workforce was only 60% the size of education's. The share of manufacturing in total employment had halved since the mid-1990s.

The international data reinforced the concern. Mortimer-Lee found that across countries, every 1 percentage point increase in the share of government consumption in GDP is associated with a 0.2 percentage point lower manufacturing share — an inverse relationship, consistent with the classical crowding-out hypothesis. The UK's manufacturing share was already 6 percentage points below where the international relationship would predict, given its level of government consumption. The UK was not merely on the wrong side of the trend. It was a pronounced outlier.

---

3. What Has Happened Since: The State at Record Size on Every Measure

The ONS Public Sector Employment bulletin for December 2025, published in March 2026, is unambiguous.

Total narrow public sector employment reached 6.19 million in December 2025 — up from approximately 5.7–5.8 million in 2022, an increase of around 400,000 or roughly 8.6% in three years. This is the headcount measure, covering the civil service, NHS, local authorities, and public corporations. It excludes the private providers of publicly-funded services that Mortimer-Lee's broader measure captures.

Within that number:

• Central government: 4.06 million — a record high. This includes schools that have converted to academy status, but the underlying trajectory exists independent of reclassifications.
• The NHS: 2.07 million — also a record high, up 17,000 year-on-year.
• The Civil Service: 555,000 — an increase of 7,000 on the year, and the ninth consecutive annual rise. The civil service is larger today than at any point in recent decades.

Public sector annual earnings growth was running at 5.9% in the year to early 2026, against 3.3% in the private sector — the state outcompeting private employers for labour at the margin, particularly skilled labour, across the entire economy.

The share of total employment accounted for by the public sector on the narrow measure stands at approximately 18% — a figure that excludes the vast swathes of publicly-funded health, care, and education provision delivered through nominally private channels.

---

4. Manufacturing Continues to Shrink

The ONS JOBS02 dataset for June 2025 records manufacturing workforce jobs at 2,555,000 — down from approximately 2.7 million in 2022 and from just over 4 million at the turn of the century. A sector that employed roughly one in eight workers in 2000 now employs fewer than one in fifteen.

The contrast with health and social work is exact. The same dataset shows human health and social work at 5,125,000 jobs in June 2025, the largest annual increase of any sector, up 68,000 year-on-year. Public administration and defence: 1,755,000. Education: 3,089,000. The three broad public-sector categories together account for just under 10 million jobs — nearly four times manufacturing's headcount.

According to Make UK's annual UK Manufacturing: The Facts 2025 report, UK manufacturing output stands at $279 billion, placing the UK 11th in global manufacturing rankings. The sector accounts for 42% of UK goods exports and 48% of all UK business R&D. These are not the numbers of a sector in managed decline — they are the numbers of a sector being systematically starved of the labour it needs to grow.

---

5. How Crowding Out Works

The mechanism is not mysterious. Labour markets clear. When the state expands hiring — especially in skilled occupations in health, administration, and education — those workers come from somewhere. The alternative career paths not taken are counterfactual and therefore invisible. But at the aggregate level, they show up in the data.

The public sector has grown its headcount by roughly 400,000 in three years on the narrow measure. Private sector earnings growth has lagged public sector earnings growth materially over the same period: 3.3% versus 5.9% in annualised terms. The state is not just hiring more — it is offering better wage growth, funded by taxation and borrowing rather than productivity. Private employers, including manufacturers competing for engineers, technicians, and graduates, are bidding against a party that cannot go broke.

This is not a new observation. Robert Bacon and Walter Eltis first formalised the argument in 1978. Mortimer-Lee's 2023 paper acknowledged the intellectual lineage. What Mortimer-Lee added was updated data showing the problem had not resolved in the intervening 45 years — it had compounded.

A 1 percentage point rise in government consumption being associated with a 0.2 percentage point fall in manufacturing's share of GDP may look like a weak relationship. But the UK sits 6 percentage points below the international trend line. That gap does not appear by accident.

---

6. The International Comparison: Extreme Outlier

The NIESR analysis compared the UK against a broad cross-section of countries. The relationship between government consumption and manufacturing share is statistically weak — individual country cases deviate substantially. Italy, Mortimer-Lee noted, has government consumption nearly 8 percentage points higher than the UK as a share of GDP, and yet its manufacturing share was 16.6% of GDP in 2019, compared with the UK's 9.7%.

The Italy comparison cuts both ways. It shows the relationship is not mechanical. But it does not rescue the UK's position. Italy is an exception. The UK is the outlier. Of the countries surveyed, only Luxembourg, Greece, Norway, Australia, and Iceland had lower manufacturing shares than the UK. These are economies shaped by specific and obvious structural factors — financial centre, geography, natural resources. The UK shares none of these extenuating circumstances. It is a large, diversified, historically industrial economy with a manufacturing share that belongs in a different list entirely.

The question is not whether the UK would automatically regain German manufacturing share by cutting public sector headcount tomorrow. It would not. The question is whether a country can sustain and grow a productive industrial base while systematically allocating an ever-larger share of its working-age population to activities funded by the state rather than by market output. The data says: apparently not, in the UK's case.

---

7. Two People, One Economy

Person 1 is doing everything right. Qualified. Experienced. Works in a manufacturing business or a private services firm. Pays taxes. Saves. Feels like the system is not rewarding effort in the way it once did. Real wages grew 0.4% in the year to early 2026, according to the ONS Labour Market Overview for March 2026. Not 0.4% above inflation — 0.4% is the inflation-adjusted figure. Their employer cannot offer the wage growth that public sector jobs can, because their employer has to generate revenue rather than draw it from taxation.

Person 2 is 21. Unemployed. Youth unemployment for 18–24-year-olds stood at 14.0% in October–December 2025 — the highest in five years, according to the same ONS release. Above the EU average for the first time since records began. Person 2 is not work-shy. The labour market has contracted around them. The entry-level jobs in hospitality and retail that historically absorbed young workers have been hit by National Insurance increases that raise the cost of hiring. Graduate employers have slowed recruitment. Manufacturing is not growing its headcount. The growing sectors — health and social work, public administration, education — typically require qualifications and experience Person 2 does not yet have, or are publicly funded positions with restricted hiring.

Person 1 is being taxed to fund a state that pays better than their employer. Person 2 cannot find work in an economy where the growing sectors are not generating the kind of entry-level private employment that would give them a start. Both outcomes are connected. Both are products of the same structural shift.

---

8. A Country That Employs People to Administer Things Rather Than Make Things

The UK in 2026 has a narrow public sector employing 6.19 million people, a record. A central government employing 4.06 million, a record. A civil service that has grown for nine consecutive years. An NHS that has grown to 2.07 million, a record. And a manufacturing sector with 2.55 million jobs — down from over 4 million at the start of the century, continuing to decline, ranked 11th globally by output but heading nowhere on employment.

The broad measure — the one Mortimer-Lee used in 2023 — would show the picture in even sharper relief. Education at 3.09 million. Health and social work at 5.13 million. Public administration at 1.76 million. Add them together and you are at roughly 9.9 million jobs in sectors that are predominantly state-funded, against 2.55 million in manufacturing. The ratio of broadly-defined public work to manufacturing has widened since 2022.

NIESR's 2023 finding was that the broad public sector had absorbed 60% of all employment growth between 1997 and 2022. There is no reason to think the subsequent three years have reversed that ratio. The employment direction across every measure points the same way: more state, less manufacturing, rising unemployment as the private sector contracts and the state cannot absorb everyone who would otherwise work in it.

The UK is not in managed industrial transition. It is not trading lower manufacturing employment for higher productivity or higher wages. Real wage growth in early 2026 is 0.4%. It is becoming, systematically and at increasing pace, a country that employs people to administer things rather than make things — and is paying the price in stagnant living standards, rising unemployment, and a youth generation locked out of a labour market that has not built a place for them.

The NIESR paper in 2023 was titled as a question: Is the UK public sector too large?

The data three years later does not leave much room for a question mark.

---