Got Softer, or Got Sicker: Britain's Dependency Rate

The April 2026 ONS labour market release tells a flattering story. Unemployment at 4.9%. Vacancies at their lowest in five years. Earnings still rising at 3.6%. On paper, a functioning economy.

The number that matters is not the unemployment rate. It is the share of working-age adults who depend on the state for part or all of their income — either because they are not in work, or because they are in work and the state is topping up their wage.

Call it the Dependency Rate. It is built from three numbers the government publishes but rarely adds together.

Add the three together and divide by the 43.4 million working-age adults in Britain. The answer is 31%.

Roughly one in three working-age Britons depends on the state for part or all of their income. The unemployment rate describes 4.9% of the country. The Dependency Rate describes the rest of it.

The 2.6 million in-work Universal Credit claimants are participating in the labour market. That is not in dispute. They produce output, they pay tax, they turn up. What the Dependency Rate captures is that they, and the people not working at all, rely on the state to make the household arithmetic work. Disengagement is one story. Dependency is the larger one.

Methodology: unemployment and inactivity are UK figures; the in-work UC caseload is Great Britain only, reflecting Northern Ireland's separate benefit system (a gap of about 2% of the total). The incapacity and disability caseload cited below is England & Wales. Normalising all three to a consistent geography changes the headline rate by less than one percentage point.

The trajectory

Through the 1980s, Britain's labour market was defined by unemployment. Thatcher's recession pushed the rate to 11.9% by 1984. From 1997 onwards, unemployment came down. Inactivity did not. The line simply shifted. People stopped being counted as looking for work and started being counted as not looking for work. The benefit bill did not fall. It changed shape.

That shape has held for twenty-nine years. Five Prime Ministers. Two financial crises. A pandemic. The architecture does not change.

The UK is now the sick man of the G7 — the only G7 economy with an employment rate lower than before the pandemic, and the highest long-term-sickness inactivity rate in the group.

Got softer, or got sicker

There is one line on the chart that is genuinely new.

In May 2019, 2.9 million working-age adults in England and Wales claimed an incapacity or disability benefit. In August 2025, the figure was 4.5 million. A 55% increase in six years. Among claimants under 25, roughly 70% of new PIP awards are now for mental or behavioural conditions. PIP claims in the 16-24 age group rose 116% between January 2019 and December 2024.

There are two possible explanations. Either the country genuinely got that much sicker between 2019 and 2025, in which case the pandemic, the state of the NHS, and the decline in population health constitute a national emergency no government has treated as one. Or the threshold for being signed off from the labour market has collapsed, in which case the state has knowingly expanded the benefit footprint to disguise long-run economic drift.

There is a defensible third reading. A public health crisis concentrated in younger cohorts. NHS waiting lists that prevent return to work. A diagnostic system now picking up conditions that would have gone unlabelled twenty years ago. None of it is invented. But each of those is itself an outcome of policy. A decade of waiting-list growth is a decision. A mental-health system that routes young adults to benefits faster than to treatment is a decision. The country getting sicker is not a defence of the system that presided over it. It is a restatement of the same legacy in different language.

The current government's own March 2025 reforms, tightening Personal Independence Payment eligibility to remove an estimated one million people from the caseload, are a de facto admission that the system has, at minimum, been too loose. If it were not, there would be nothing to tighten.

Either the country got sicker, or the system got softer. Both answers require a rebuild.

Name a country

Britain's welfare spend as a share of GDP is near the top of the OECD. Its working-age participation rate is not. Germany, the Netherlands, Switzerland and most of the Nordic bloc spend similar or more on welfare, and have higher labour market participation. The combination of a large welfare state and weak participation is not a paradox. It is a policy outcome.

The arithmetic is starker at the household level. A full-time worker on the National Living Wage takes home about £22,500 after tax. An economically inactive claimant on Universal Credit with average housing benefit and standard-rate PIP receives around £25,000. On the highest PIP rate, £27,500. For an out-of-work single parent claiming PIP for anxiety plus benefits for a child with ADHD, it can reach £36,900 — more than £14,000 above the after-tax wages of a full-time worker on the legal minimum. (Source: Centre for Social Justice, 2025.)

It was assembled in instalments. Blair's tax credit regime in the late 1990s. The post-2010 preservation of its architecture despite different rhetoric. The pandemic-era expansion of sickness-based inactivity. A Universal Credit system that, by design, subsidises the wages of people whose employers could pay them properly.

No single decision produced a country in which nearly a third of working-age adults depend on the state for part or all of their income. A series of decisions did, spread across five Prime Ministers and every government that has held office since 1997.

The legacy

This government may fall this week. When it does, its true economic legacy will not be the unemployment rate. It will be the one-in-three figure that the unemployment rate was designed to conceal.

Whatever replaces the current administration will inherit this arithmetic intact. The first honest sentence any successor can say is the one this government spent its term avoiding.

Either the country got sicker, or the system got softer. Both answers require a rebuild.