# Fessing up: the UK is more productive than they thought _The ONS now says its own labour survey has been understating how productive Britain is. On the better data, output per hour has grown twice as fast since 2019 as the number used to set tax and interest rates._ Neil Woodford · 24 August 2026 · 4 min read ![Street art arrow with "better days ahead" written inside it](https://cdn.sanity.io/images/v3acfbvo/production/8636917e314683d198178fcba24f63cae1968f80-2880x1620.jpg?w=1600&fit=max&auto=format) --- I have never had much time for economic and financial market convention. That is not non-conformism for its own sake. It comes from wanting to understand what is actually going on in a complex and fast-changing world. My view is that without clarity about what is knowable, policymakers are condemned to compound their errors, and the rest of us live with the sub-optimal outcomes. In some important respects that is precisely what has been going on in the UK, for far too long. I first wrote about the ONS’s measurement problems in [November 2024](https://www.noisecancelling.co/read/lies-damned-lies-and-misleading-data), and about the reliability of its Labour Force Survey (LFS) in particular [the following March](https://www.noisecancelling.co/read/data-vs-common-sense-fight). At the time it may have looked like a slightly esoteric rabbit hole. I thought it was significant then. I believe it even more now. _(Neil in the margin: The LFS is a household survey. The ONS asks a sample of households who is in work and for how long, then grosses the answers up to the whole country. Response rates have fallen sharply since the pandemic, which is where the trouble starts.)_ ## The number underneath everything The ONS has used the LFS as its key measure of how many people are in work in the UK and how many hours they have worked. That is important enough as a read on the health of the labour market. It matters more than that, because the same series is the denominator in the calculation of the UK’s much-maligned productivity performance. The OBR and the Bank of England, who between them drive the UK’s fiscal and monetary policy decisions, have both used LFS data in their productivity calculations, historically and prospectively. For the OBR, the productivity outlook is the key variable in the models that dictate how fast the economy will grow, and, in turn, what the OBR “suggests” the government needs to do with spending, borrowing and tax to meet its fiscal targets. For the Bank of England, the productivity forecast sets what it sees as the economy’s “speed limit”: how fast the economy can grow without generating excess inflation. Weak productivity forecasts lower that speed limit, stronger ones raise it. So where productivity is forecast to grow slowly, the institutional models say that even modest growth can trigger higher inflation, which in turn argues for higher interest rates. The outlook for productivity is therefore a crucial variable in the setting of both fiscal and monetary policy. It could barely be more important. And the data on which it relies has been faulty, as I first pointed out nearly two years ago. Because of failures in the reliability of the surveys underneath it, the LFS has consistently overstated the number of people in work and the hours they have worked. Overstate the labour input in the productivity equation, and you depress this key measure of the UK’s economic performance, and, most importantly, you depress the forecasts of potential output growth that follow from it. It was this flawed data that fed the OBR’s [November 2025 downgrade](https://obr.uk/efo/economic-and-fiscal-outlook-november-2025/) to its medium-term productivity forecast, from 1.3% to 1%. That not only brought the OBR more into line with the Bank of England’s customary downbeat expectations, but it also drove the medium-term growth forecast for the economy down to 1.5%, which added to the fiscal pressure bearing down on the government and fed the “black hole in the nation’s finances” narrative. _The background:_ [Dog with a bone: UK productivity mismeasurement](https://www.noisecancelling.co/read/dog-with-a-bone-uk-productivity-mismeasurement) — I’ve said it before, but it bears repeating — the ONS’s productivity data simply doesn’t make sense. According to its latest figures, UK manufacturers are hiring more people to produce less, and the labour market is supposedly booming while productivity stagnates. None of this aligns with reality. The data is broken, yet it remains the foundation for critical economic forecasts and policy decisions. ## The admission On 18 August the ONS [published its latest productivity figures](https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/labourproductivity/articles/ukproductivityintroduction/apriltojune2026andjanuarytomarch2026) and, in effect, owned up again. For the time being, it says, HMRC’s Real Time Information data should be used to assess, you guessed it, labour productivity. _(Neil in the margin: RTI comes from HMRC’s payroll records rather than from a survey. It is close to a census of payrolled employees, with the self-employed added from self-assessment returns, so falling response rates cannot distort it.)_ _[Embedded media](https://www.theguardian.com/business/2026/aug/16/uk-productivity-data-rachel-reeves-national-statistician)_ The difference is not a rounding error. In the LFS-based data, output per worker rose by 0.4% in the year to Q2 2026, while output per hour fell by 0.2%. On the HMRC data the ONS now prefers, output per worker rose by 1.4%, a full percentage point higher, and output per hour rose by 0.7% instead of falling. Perhaps more shocking is the longer view. Measured against its 2019 average, the LFS-based series shows output per hour up 2.3% by Q2 2026. The more accurate HMRC-based series shows the same measure up twice as much, at 4.6%. As I have been wittering on about for some time, this all comes back to the LFS overstating the number of hours worked in the UK. My favourite economist drew my attention to the flaw more than two years ago. It has taken until now to emerge as a real issue for the economic consensus. ## Not the ONS alone _[Embedded media](https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/labourproductivity/articles/ukproductivityintroduction/apriltojune2026andjanuarytomarch2026)_ In July, the LSE’s [Centre for Economic Performance](https://cep.lse.ac.uk/pubs/download/occasional/op074.pdf) arrived at the same place from the other direction. Working from the administrative data, Niki Barbas, Anna Valero and John Van Reenen find annualised productivity growth of 1.6% between Q3 2024 and Q1 2026, against 0.3% over the decade before it. Even on their most conservative measure, that is what they call a “meaningful productivity pickup”. What is causing this inflection is the subject of some debate. What is not in debate is that the monetary and fiscal authorities in the UK need to address the mismeasurement and stop forecasting the economy from data their own statisticians no longer stand behind. They should also say plainly what follows from it, as I have been saying for some time: the UK’s productivity performance has been significantly better than they assumed, which makes their assumptions about future growth and inflation both too pessimistic and wrong.