A rebound that still owes the country five per cent

For fifteen years the United Kingdom has been a country in which output per hour refused to grow in the way a rich economy is supposed to grow. Wages noticed. Politics noticed, then found other things to shout about. This summer a quieter claim arrived: that the two years to the second quarter of 2026 have seen output per hour rise at about 1.1 per cent a year on the more trustworthy measures, after a stretch in which it fell. That is faster than the late 2010s. It is also, if the Resolution Foundation’s arithmetic is right, a country still sitting about 5 per cent below the pre-pandemic path — on the order of £150 billion of lost output a year, or £4,500 per worker.

The Labour Force Survey, which has spent the decade tripping over its own sample, still points the other way: roughly minus 0.2 per cent a year on the same window. Workforce Jobs and tax-administrative employment both say plus 1.1 per cent. When a household survey and the tax system disagree about how many people are working, believe the tax system. The Office for Budget Responsibility has already marked its medium-term productivity assumption down toward 1 per cent. For once the official forecast and the better data are in the same postcode. That is not the same as being on the old path.

A rebound is not a restitution

The 2020s delivered a sequence of level hits: the pandemic, Brexit’s non-tariff thickening, the energy shock after Russia’s full-scale invasion of Ukraine, and now a further energy disturbance from the Middle East. A country can grow at a respectable rate from a permanently lower path. Commentators will call that a renaissance. It is not. It is a scar with a pulse.

One per cent a year is not the 2000s. It is also no longer zero. Both sentences can be true.

What might be driving the pickup? Sector mix is the boring candidate and should be checked first: if hours have shifted toward higher-output activities, the aggregate can rise without any given worker becoming more effective. Capital deepening is the candidate everyone wants, especially if firms have finally been forced to substitute machines for labour they cannot hire. Measurement is the candidate nobody wants: the LFS is not the only series that can be wrong.

Do not annex this to an AI story without evidence

There is a fashion for attributing any pleasant productivity print to large language models. The fashion should be resisted until someone shows the hours and the output in the sectors that actually use the tools. Memory-chip prices, which the Bank thinks will add to goods inflation this year, are a reminder that the AI boom is at present a demand shock for hardware as much as a supply shock for services. A demand shock for hardware can raise measured productivity in a handful of plants and leave a GP surgery untouched.

The policy implication is unfashionable. If the rebound is real, the OBR’s caution looks less like masochism and more like adult supervision, and the Budget’s tax take is being asked to fund a state whose trend growth is 1 per cent, not 2. If the rebound is a statistical truce between broken surveys, then the fiscal arithmetic is worse than the Chancellor has admitted. Either way, this is not a moment for a press release about dynamism. It is a moment to keep reading the tax data, to stop treating the LFS as a thermometer, and to remember that a country can clap for 1.1 per cent and still be poorer than it should have been.

This essay is commentary, not advice. Sources: Bank of England, Office for National Statistics, House of Commons Library, Office for Budget Responsibility, Resolution Foundation. See the editorial method.