Over the past two years, UK output per hour rose 1.1%. Over the same two years, it fell 0.2%.
The first figure is Resolution Foundation analysis, reported on 25 August under the headline that productivity rises despite official figures. The second is the official series. The gap between them is 1.3 percentage points on the measure that determines, more than any other, whether a country gets richer.
Why Two Measures of the Same Thing Differ
Productivity is not observed. It is constructed by dividing one estimate by another.
The numerator is output, which for large parts of the economy has no market price and must be imputed. The denominator is hours worked, which comes from surveys of employers and households that disagree with each other and are revised repeatedly.
Small differences in either produce large differences in the ratio, and the differences compound when the underlying growth rate is close to zero. When productivity is genuinely growing at 2% a year, a measurement disagreement of a few tenths is noise. When the true rate is somewhere between minus a fraction and plus a fraction, the same disagreement determines the sign.
That is the position Britain is in, and it is why a think tank and a statistical agency can look at the same economy and disagree about direction rather than magnitude.
The Explanation Offered
The Resolution Foundation’s account of its own result is notable for what it rules out.
The improvement, it argues, reflects stronger output from employees already in existing jobs rather than an AI-driven productivity boom.
That is a claim worth registering, because the default explanation for any productivity improvement in the current climate is AI, and here is an organisation with a positive productivity finding declining to attribute it to that. It also sits comfortably alongside Bank of England research finding the clearest AI-linked gains in the industries that build AI rather than the ones using it.
Two independent analyses, using different methods, both landing on the conclusion that the AI story is not yet where the measurable movement is, is more informative than either alone.
The Hiring Data Points the Same Way
The same day’s figures contain a labour market picture that fits the think tank’s explanation, and complicates how comfortable anyone should be about it.
Total UK vacancies stood at 791,490 in July, down 9.6% on the year. Non-graduate entry-level vacancies were 192,864, down 8%. Graduate vacancies were 8,383, down 45.6%.
That last figure is extraordinary in context. Graduate roles fell at roughly five times the rate of the market as a whole, and now number fewer than nine thousand across the country.
Set that beside a productivity gain attributed to existing employees producing more, and a mechanism suggests itself. If output holds up while firms stop adding people, output per hour rises. That is arithmetically a productivity improvement regardless of what caused it.
It is also a description of a workforce absorbing more work rather than a workforce equipped to do work better, and those have very different implications for whether the gain persists.
Policy Will Push the Number the Same Way
There is a further reason to expect measured productivity to improve over the next two years, and it is not a happy one.
The government’s own impact assessment of the zero-hours contract reforms puts the additional cost to employers at up to £3 billion a year. CIPD research found 65% of employers using such contracts expect higher management and HR costs, and 31% believe redundancies could result.
Whatever one thinks of the reforms on their merits, the arithmetic is straightforward. If employers respond by reducing hours, and output does not fall by as much, output per hour rises. A productivity improvement produced that way is indistinguishable in the statistics from one produced by better equipment, better management or better technology.
This is a long-standing awkwardness in the measure. Labour productivity improves when the workforce shrinks faster than output, which is why productivity often looks strongest in the aftermath of recessions. The number rewards shedding low-output work regardless of what happens to the people doing it.
Adding it together: hiring is already falling sharply, a policy change with a multi-billion pound cost estimate is arriving, and a think tank has found output per hour rising for reasons it attributes to existing staff. Each of those is separately reported as either good or bad news. They may all be the same phenomenon, described from different angles.
That does not make the Resolution Foundation wrong. It means a rising productivity number over the next two years will need its cause established rather than assumed, and the cause will not be visible in the headline figure.
Which Number Should a Business Use
For a firm making a decision, the honest answer is that neither figure should carry much weight on its own.
National productivity is an aggregate across millions of businesses in wildly different conditions. It tells a manufacturer very little about manufacturing and a law firm nothing about legal services. Sectoral figures are more useful, and a firm’s own output per hour is more useful still, because it is measured rather than estimated.
Where the national number matters is indirectly, through policy. The Office for Budget Responsibility’s productivity assumption drives the fiscal forecast, which drives the tax and spending decisions that reach every business. A dispute about whether productivity is rising or falling is therefore a dispute about the tax environment two years out, conducted in a language most firms never read.
That is the practical reason to pay attention to a 1.3 point measurement gap: not because it changes what any company should do this quarter, but because the resolution of it shapes the budget arithmetic.
Why the Disagreement Matters More Now
The context given for the Resolution Foundation work is that it accompanies data showing the UK among the fastest-growing G7 economies in the first half of the year.
Growth and productivity are different things, and the relationship between them is the whole question. An economy can grow by adding workers and hours, which raises output without raising output per hour and does not raise living standards per person. Or it can grow by producing more from the same inputs, which does.
If the official series is right and productivity fell 0.2%, then the growth is the first kind, and it is not the sort that compounds. If the Resolution Foundation is right, some of it is the second kind, and the outlook is materially better.
The same growth figures support both readings, which is a good illustration of why the productivity measure is the one that matters and the one nobody can agree on.
What Would Settle It
Disputes of this kind get resolved slowly and quietly, through revisions rather than argument.
Official productivity estimates are revised as better source data arrives, and the revisions on a two-year window are routinely larger than 1.3 points. It is entirely possible that in a year’s time the official series will show the period as positive, at which point the disagreement will have evaporated without anyone conceding anything.
Until then, the defensible position is that UK productivity over the past two years was approximately flat, with a plausible range spanning both published estimates. That is less satisfying than a direction, and it is what the evidence supports.
It is also worth holding both numbers rather than picking the congenial one. The temptation with a contested statistic is to adopt whichever version fits the argument being made, and the presence of two credible estimates makes that easier rather than harder. The more useful discipline, here as with any dataset that admits more than one reading, is to state the range and say plainly which parts of it the evidence cannot yet rule out.
About The Author
More Stories
Waste Incineration Was About to Cost £48 a Tonne More
Lucky Egg Grew Sales 515% to Top Britain’s Growth 100
Extreme Heat Starts Costing Retailers at 25.1C