# The great UK productivity myth _The ONS now admits its own labour survey understated Britain's productivity by half. The full story: how the establishment got it wrong, what it means for interest rates and tax – and the dated record of every time we said so._ Neil Woodford · 28 August 2026 · 14 min read ![A traffic marshal in hi-vis holds a lollipop sign reading STOP WORK on a London street](https://cdn.sanity.io/images/v3acfbvo/production/ea23357708befea4c5561f3a6b92ae17fb02fa8e-1944x2592.jpg?w=1600&fit=max&auto=format) _Photo: Mason C. / Unsplash_ --- I should apologise upfront for bringing this subject up again, but I hope that those who make it to the end of this note will understand why, and forgive me for doing so. I am not sure I have written about anything so important in the more than two years I have been doing this, and by now regular readers will know that I have a proper bee in my bonnet about it. The issue that has got under my skin again is the great productivity debate, which sits at the heart of the UK’s monetary and fiscal policymaking and, consequently, could not be more critical to the economic outlook. I wrote [a note about this earlier this week](https://www.noisecancelling.co/read/fessing-up-uk-productivity), but I want to do justice to the full scale of what I see as a monumental establishment balls-up encompassing the ONS, the Bank of England, the OBR and, finally, the tax and spending decisions of the incumbent government. _(Neil in the margin: The plumbing of UK economic policy: the ONS (Office for National Statistics) produces the data, the Bank of England sets interest rates off it, and the OBR (Office for Budget Responsibility) uses it to score the government's fiscal room for manoeuvre. Get the data wrong and all three misfire.)_ _The brief note I wrote earlier in the week:_ [Fessing up: the UK is more productive than they thought](https://www.noisecancelling.co/read/fessing-up-uk-productivity) — 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. The consensus (incorrect) view on this subject has also fed countless pages of doom-laden financial media commentary, which, of course, helps drive the widespread view that the UK economy is disabled, underperforming, unable to grow, and confronting some kind of catastrophic denouement. To give the subject the attention it deserves, I want to address it properly from start to finish, including a brief introduction covering what productivity is, what it measures, what economic theory says about its importance and how it drives policy decisions. To keep it entertaining, I will use as many graphics as possible to help tell the story. I promise there will be some surprises, even for readers familiar with this issue. So first to the theory bit. If you’d rather watch, this week’s show covers the same topic: _[Watch: We discussed this topic on The Show — Britain’s Productivity Crisis: A Great Deception?](https://www.noisecancelling.co/the-show)_ ## What is productivity, and how is it measured? Productivity is the measure of how efficiently economic inputs – labour, capital and technology – are converted into outputs of goods and services. It is the primary long-term driver of economic growth, rising real wages and living standards. At an economy level, it is measured by reference to output per worker or, more conventionally, output per hour worked: GDP divided by the total hours worked in the economy. ![Figure](https://cdn.sanity.io/images/v3acfbvo/production/06a93255a741b9313d7e02cc353da22659371d0e-1536x1024.png?w=1600&fit=max&auto=format) ## Why is productivity so important? If you got this far, you already know: it is the primary long-term driver of economic growth, rising real wages and living standards. But how does it drive interest rates and taxes? ### Monetary policy First, monetary policy. In its official literature on productivity, the Bank of England offers some conflicting messages, which are not that helpful but are, not surprisingly, a product of the detached-from-reality economic theory the institution is so fond of. On the one hand, the Bank suggests that a temporary, one-off increase in productivity lowers production costs, raises output capacity and so exerts downward pressure on prices. It then goes on to say that a sustained increase in productivity growth raises household incomes, boosts spending and investment, and can therefore raise prices if demand outpaces supply growth. I get this in theory, but in the real world, households and businesses do not anticipate economy-wide increases in productivity, however neatly they might in a complicated theoretical model. Yet the Bank also tells us the opposite: that weak productivity keeps inflation alive. As its [January 2020 Monetary Policy Report](https://www.bankofengland.co.uk/monetary-policy-report/2020/january-2020/in-focus-supply-and-spare-capacity) put it: “Weaker potential supply growth reduces the pace of GDP growth that is consistent with the MPC meeting its 2% inflation target — it acts as a ‘speed limit’ on the economy.” _So, which is it then?_ My view is that productivity growth is, by definition, an increase in potential supply, so better productivity growth raises the economy’s non-inflationary speed limit. This is the more widely held view, and it seems to be what the Bank of England really believes. It is this approach which significantly drives its interest rate decisions, its view of spare capacity in the economy and its estimate of the neutral real interest rate, or r* – the rate that neither stimulates the economy nor holds it back. _(Neil in the margin: r* is a theoretical, unobservable rate — you can only estimate it. Because it depends on the economy's supply potential, understated productivity would drag the Bank's r* estimate too low, and with it its whole sense of how restrictive current rates really are.)_ ### Fiscal policy As far as fiscal policy is concerned, the OBR views its judgement on medium-term productivity growth as one of the single most important drivers of its entire economic and fiscal outlook. A good example was the OBR's recent downgrade of its medium-term productivity growth forecast from 1.3% to 1%, which produced a lower medium-term growth forecast and, in turn, a more challenging fiscal position at the end of this parliament. That led Rachel Reeves to increase taxes again in last November's budget. _(Neil in the margin: A 0.3-point cut sounds trivial, but compounded over the forecast horizon it lowers the entire projected level of GDP — and hence tax receipts — enough to blow a hole in the fiscal rules. That is the mechanical link to Reeves's tax rises.)_ Productivity, and how the UK’s monetary and fiscal authorities view it, could barely be more important. In summary, it is the key driver of interest rate and fiscal policy decisions. ## What went wrong at the ONS Having outlined the background, it is time to explain what has gone wrong in measuring this critical variable and how, as a result of prolonged mismeasurement, major policy errors have occurred. Nearly two years ago, I started writing about the problems the ONS was having measuring how many people are in work and, as a result, how many hours have been worked in the economy. These problems were first drawn to my attention by my favourite economist, who spends a lot of time poring over UK economic data and unearthing gems like this. We [publish his monthly economic bulletins here on Noise Cancelling](https://www.noisecancelling.co/topics/economic-updates). The mismeasurement was and still is significant, not just because it gave a misleading picture of the health of the labour market but, arguably more importantly, because total hours worked is the denominator in the calculation of UK productivity. In summary: the data behind the ONS's measure of this critical variable comes from a survey whose number of respondents has fallen significantly since the pandemic. As a result, the survey was overestimating the number of people employed in the economy, and with it the number of hours worked. _(Neil in the margin: This is the Labour Force Survey. Post-pandemic response rates collapsed, so the sample became less representative — a smaller, more self-selecting group being grossed up to national totals, which is exactly how a survey drifts away from reality.)_ The anomaly was revealed in part because alternative labour market measures compiled by HMRC gave a very different picture. My view nearly two years ago, as now, was that the ONS data was overstating hours worked, that the HMRC data was more reliable – they should know how many people are paying tax – and that the knock-on effect was an understatement of the UK's productivity performance. _(Neil in the margin: This is the PAYE Real Time Information feed — an administrative count of everyone on a payroll, captured each time employers run wages. It isn't a sample, so it sidesteps the response-rate problem that undoes the survey.)_ Bit by bit, the ONS has admitted these errors, which appear to have started from mid-2023 onwards, culminating in a statement published on 18 August which has received nowhere near enough attention. In it, the ONS states: > “We are developing a new component-based approach to measuring labour productivity to address challenges associated with labour market measurement and to align with international best practice. The new approach produces estimates that are more closely aligned with Real Time Information (RTI) based measures of employment. Further details will be available in a forthcoming methods article that will be published in September.” > > — [Office for National Statistics](https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/labourproductivity/articles/ukproductivityintroduction/apriltojune2026andjanuarytomarch2026) In plain English, the ONS is acknowledging that **overstating the number of hours worked** in the economy through the Labour Force Survey has **produced understated measures of productivity**, and that it has reworked the data using the more reliable HMRC (RTI) data. That reworked data gives rise to a significantly different picture of the UK's productivity in recent years. ![All the Productivity We Cannot See](https://r4at4qm6kmohrtvq.public.blob.vercel-storage.com/charts/chartDoc.prodmyth.c1-oph-lfs-vs-rti-25cd1f2f896f-light.png) _The two series told broadly the same story until mid-2020, then parted company. The gap at Q2 2026 – 104.6 against 102.2 – is the productivity the survey couldn't see._ 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. 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%. _[Chart: Twice the growth they thought — Same economy, same output. The only difference is who you ask about hours worked: a survey ever fewer people answer, or the tax records nearly everyone is in.]_ ## The outside world catches up Interestingly, this completely revised picture of the UK's recent productivity performance has coincided with work from outside the statistics office pointing the same way, focused on a more recent inflection in this critical indicator. The first came in July from the LSE's Centre for Economic Performance – [a report](https://cep.lse.ac.uk/pubs/download/occasional/op074.pdf) put together by a group including two economists who have advised Rachel Reeves – showing annualised productivity growth of 1.6% between Q3 2024 and Q1 2026, against 0.3% a year over the preceding decade. _[Embedded media](https://www.bloomberg.com/news/articles/2026-08-26/britain-s-becoming-more-efficient-and-it-s-little-to-do-with-ai)_ The second report [was produced this week by the Bloomberg Economics team](https://www.bloomberg.com/news/articles/2026-08-26/britain-s-becoming-more-efficient-and-it-s-little-to-do-with-ai). It calculates that since the first quarter of 2025, year-on-year productivity growth has averaged around 2%, once again using a more reliable measure of employment than the official headline figures. It reports the improvement as broad-based, seen in twelve of nineteen sectors of the economy. Interestingly, the cause of this improvement is something the authors are unclear about. They see AI's impact as limited so far, while two economists from the [Resolution Foundation](https://www.resolutionfoundation.org/publications/the-macroeconomic-policy-outlook-q3-2026/) have said that while they believe the pick-up is genuine, the revival of productivity looks “as mysterious as its fall”. Before commenting on what I think is going on, it is important to compare these numbers with the assumptions embedded in the official view. The OBR assumes trend productivity grew by just 0.3% in 2024 and 0.7% in 2025, and does not expect it to reach 1% until 2030. The Bank of England, for its part, describes a prolonged productivity slowdown running all the way back to the 2008 financial crisis; [February's Monetary Policy Report](https://www.bankofengland.co.uk/monetary-policy-report/2026/february-2026) notes that productivity growth “has been subdued relative to its assumed long-run trend growth rate of 1%, averaging 0.5% per year since 2024”. _(Neil in the margin: The MPR is the Bank's quarterly set-piece laying out its forecasts and the reasoning behind rate decisions — so a productivity misjudgement embedded here propagates straight into policy.)_ In both cases, remember that these judgements rest on the flawed LFS data, which has been giving a misleading picture of the UK's labour market since about the middle of 2023. Whether these two profoundly important institutions believe the LSE and the Resolution Foundation or not, what they do need to recognise is that the data they have been basing their calculations and decisions on is flawed, and has understated the UK's productivity performance since the pandemic – according to the ONS itself, by a factor of two. ## The international comparison As a last point in this section, I wanted to look at international comparisons to put the UK's performance in context. Before showing the data, remember that the UK's productivity track record – on the flawed data – has generally been viewed as dire, indicative of the economic malaise supposedly crippling the economy. So dire, in fact, that I had come to expect international comparisons would be embarrassing. Interestingly, even on the flawed data, you can’t draw that conclusion. Based on the OBR's own numbers, although it is correct that productivity growth was nowhere near as good as in the US, the UK is not an outlier against the rest of the G7 over the last twenty years. _(Neil in the margin: The G7 — US, UK, Germany, France, Italy, Japan and Canada — is the standard peer group for this kind of comparison. The point is that even on the flawed numbers, Britain sits mid-pack rather than bottom.)_ ![Even on the flawed data, the UK is not the outlier you were told.](https://r4at4qm6kmohrtvq.public.blob.vercel-storage.com/charts/chartDoc.prodmyth.c3-g7-race-0aaeb99c4b54-light.png) _Even on the flawed LFS data, the UK travels with the pack for twenty years – only the US breaks away, and only after 2020. Nowhere near the outlier the doom narrative requires._ Indeed, had the OBR not mistakenly downgraded its productivity forecast last November – a decision based on the flawed LFS data – the UK's medium-term forecast would be second only to the US in this comparator group. If the LSE is to be believed, the UK might even top this particular league table! ![What November's downgrade did to the UK's ranking](https://r4at4qm6kmohrtvq.public.blob.vercel-storage.com/charts/chartDoc.prodmyth.c4-intl-forecasts-b65a8912a2a9-light.png) _Reverse November's mistaken downgrade and the UK's official medium-term forecast is second only to the US. On the LSE's numbers it might top the table._ As an aside: in [the material published alongside the downgrade](https://obr.uk/efo/economic-and-fiscal-outlook-november-2025/), the OBR went to great lengths to explain why it had acted. If you will excuse me, I am going to look again, briefly, at that fairy tale masquerading as economic analysis. Across the 203 pages of the November 2025 Economic and Fiscal Outlook, the OBR refers frequently to the UK's disappointing productivity performance and attributes it to a broad range of things: 1. “This ongoing weakness makes it less likely that a substantial and rapid productivity growth rebound, as seen after previous shocks, will now materialise as the Covid and energy price shocks retreat into history.” 2. “Rather, it suggests persistent weakness in productivity growth relative to the pre-financial crisis period is more likely to reflect underlying structural trends.” 3. Declining trade intensity following a resurgence in protectionism. 4. The enduring effects of Brexit. 5. Sectoral issues. 6. AI providing a smaller boost to productivity growth than the ICT revolution did before 2008. 7. Increased employment in the less productive public sector. 8. An ageing population and fewer graduates. 9. Climate change and net zero. Some of this may be true, but trying to explain why the OBR mistakenly believes productivity is so poor by invoking this catch-all list of culprits now looks embarrassing. The UK's productivity performance was not as bad as the OBR assumed, rendering this fantasy masquerading as analysis completely redundant. ## What I think is actually going on So, it’s time to lay out what I think is going on with UK productivity. This is not something I dreamt up this week: I have been writing about it consistently for nearly two years, starting with the ONS's measurement problems in November 2024 and the reliability of the Labour Force Survey the following March. The [full paper trail](#the-paper-trail) is at the end of this note. First, measuring these things contemporaneously is not easy. Measuring GDP is genuinely hard, especially the way the ONS goes about it. But measuring the number of people in work should not be, given that we are all paying tax and HMRC has a pretty good grip on that for obvious reasons. Definitive conclusions about the real causes of this productivity inflection will emerge over time. Right now, though, I can make an educated guess. There are two things to explain. The first is the easier one: productivity has been under-recorded for at least two years because of the flaws in the LFS data. Based on what the ONS is now saying, UK productivity has increased by 4.6% in total since its pre-pandemic level, not the 2.3% previously claimed. The second – the causes of the apparent productivity inflection since 2024 that the LSE has identified – is harder. I think two factors are driving it: one concerns the labour market, the other is a product of technology change, and I believe they are connected. ### The labour market Following Labour's election victory in the summer of 2024, businesses rightly anticipated that the cost of employing people would increase. A £25bn increase in employers' National Insurance and significant increases in the minimum wage – the apprentice rate alone rose 18% – alongside later changes to employment rights did exactly that. The effect is visible in the labour market data (not the LFS), in the collapse in vacancies and in the continued fall in graduate recruitment. _(Neil in the margin: Employer NICs are a tax on hiring, paid on top of wages. Raising them makes each worker dearer, which — as Neil argues — nudges firms to substitute capital for labour, mechanically lifting output per hour.)_ Increasing the cost of employment led, unsurprisingly, to lower demand for labour. It also coincided with higher business investment, as companies substituted one for the other. ![Investment is back near its 1989 highs](https://r4at4qm6kmohrtvq.public.blob.vercel-storage.com/charts/chartDoc.prodmyth.c5-investment-share-4fbb3748b0cd-light.png) _Investment is close to its highest share of GDP since 1990. Chronically underinvesting economies do not look like this._ The obvious net result is higher productivity. ### The AI industrial revolution The other factor at play is, I believe, the early deployment of AI across the business services sector – which, unlike the OBR, I believe will have a very significant impact on productivity, as will, in time, the deployment of humanoid robots. _[Embedded media](https://www.youtube.com/watch?v=RZMj7C41AuQ)_ So the inflection in UK productivity growth since 2024, which many economists now recognise, is the product of a relative shift in the price of labour and capital, combined with the early impact of AI. It is that simple – and it has absolutely nothing to do with Brexit, trade, climate change, net zero or an ageing population. ## What should happen now? Before I conclude this tome, I should say what I think might happen after this fundamental reappraisal of the UK's productivity performance. The first thing is that it should lead to a recalculation of the Bank of England's assessment of the economy's speed limit – the maximum rate at which it can grow without generating inflation – and of its medium-term growth expectations. In my opinion, it should prompt a re-evaluation of the appropriateness of current interest rates, and it should increase the pressure on the committee to cut later this year. As for the OBR, it should lead to a complete reboot of its UK growth forecasts, which are clearly too low. With the correct productivity assumptions in the model, I suspect a rather better fiscal position would be revealed, and there should naturally be no need to increase taxes further – although I suspect a Burnham-led government could not resist the temptation to use any extra fiscal headroom to grow spending even more. ## In summary 1. As a result of a serious mismeasurement of the number of people in work and the total hours worked in the economy, UK productivity has been seriously underestimated. 2. This must have a profound impact on the Bank of England's and the OBR's understanding of the economy's recent history, and on their forecasts for inflation, growth, interest rates and fiscal policy. 3. The consensus doom-laden narrative on the UK economy should be re-evaluated. In fact, the UK economy could be on the cusp of a period of stronger growth accompanied by low inflation and lower interest rates. Now that really would be something to celebrate. ## The paper trail One last thing. When an establishment consensus collapses, everyone turns out to have known all along. So, for the record, here is the case as I made it, in order, with what I said at the time. [**20 November 2024 · Lies, Damned Lies, and Misleading Data**](https://www.noisecancelling.co/read/lies-damned-lies-and-misleading-data) — where it started. Hours worked up about 1% in Q3, GDP up 0.1%: one of them had to be wrong. > Hours worked are up, but output is flat—therefore, productivity must have fallen. But I would suggest that businesses don’t pay people more to produce less. This isn’t a sign of economic malaise; it’s a measurement problem. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/lies-damned-lies-and-misleading-data) [**17 March 2025 · Data vs Common Sense: Fight!**](https://www.noisecancelling.co/read/data-vs-common-sense-fight) — the Labour Force Survey named for the first time. > In recent months, the ONS has admitted problems with several other important datasets, including the labour force survey, earnings data, and its living costs and food surveys—all critical components used to calculate official inflation and GDP figures. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/data-vs-common-sense-fight) [**17 June 2025 · What a softer UK labour market really tells us**](https://www.noisecancelling.co/read/what-a-softer-uk-labour-market-really-tells-us) — the LFS set against the payroll data for the first time (+90,000 on the survey, −100,000 on the payroll), and the conclusion stated plainly. > I am pretty confident that productivity has been understated in the UK for reasons I explained in Data vs Common Sense, but nevertheless, it is often cited as a critical weakness in the economy. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/what-a-softer-uk-labour-market-really-tells-us) [**16 September 2025 · Uneducated educated guesses**](https://www.noisecancelling.co/read/uneducated-educated-guesses) — on the OBR's forecasting, two months before the downgrade. > Finally, it would seem very odd to expect that the AI industrial revolution unfolding right now will bypass the UK economy. It will undoubtedly have a very positive impact on both public and private sector productivity over the next five years. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/uneducated-educated-guesses) [**8 October 2025 · You couldn't make it up**](https://www.noisecancelling.co/read/you-couldnt-make-it-up) — the OBR caught marking productivity down and up at the same time. > So, just as the OBR is about to tell the Chancellor that lower productivity in the next five years will mean that she will have to increase taxes, it is being forced to upgrade its contemporaneous calculation of UK productivity. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/you-couldnt-make-it-up) That same month, the argument went on film – ten months before the ONS conceded the point. _[Embedded media](https://www.youtube.com/watch?v=u4vyNaJg60s)_ [**21 October 2025 · Dog with a bone: UK productivity mismeasurement**](https://www.noisecancelling.co/read/dog-with-a-bone-uk-productivity-mismeasurement) — the full LFS-versus-HMRC case, and the prediction that the ONS itself would eventually disown the numbers. > I think that the ONS’s productivity measurements will, sooner or later, be revealed as being in this growing group of survey-based data series that cannot be relied upon… > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/dog-with-a-bone-uk-productivity-mismeasurement) _The flagship piece:_ [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. [**14 November 2025 · Roundup of the week**](https://www.noisecancelling.co/read/roundup-of-the-week-14-november-2025) — against the MPC's own suggestion that the survey was understating employment. > My diagnosis of the LFS is the opposite. In my judgement, it is overstating the number of people in work, as evidenced by a comparison between the LFS and the more reliable HMRC data represented by the Real Time Indicators series. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/roundup-of-the-week-14-november-2025) [**19 November 2025 · Punch and Judy Show**](https://www.noisecancelling.co/read/punch-and-judy-show) — the mechanism, spelled out. > Remember as well that the ONS has itself cast considerable doubt on the accuracy of its employment data (the Labour Force Survey), which it in effect believes is overstating the number of people in work (this is the denominator in the productivity calculation). > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/punch-and-judy-show) [**24 November 2025 · Pre-Budget Politics: Lies, Lunacy and a £20bn Mistake**](https://www.noisecancelling.co/read/pre-budget-politics-lies-lunacy-20bn-mistake) — days before taxes were raised on the strength of the forecast. > My guess is that this process will be seen as flawed as it becomes increasingly clear, over the next few years, that UK growth outcomes are far better than the OBR has predicted, which in turn will drive better productivity and better tax revenues. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/pre-budget-politics-lies-lunacy-20bn-mistake) [**20 February 2026 · Roundup of the week**](https://www.noisecancelling.co/read/roundup-of-the-week-20-february-2026) — the quantified call, six months before the ONS's own revision said the same thing. > The LFS survey shows employment increased in 2025 by 380,000, or a little over 1%. HMRC data, on the other hand, showed employment falling by 120,000 — a difference, in total, of 500,000. Importantly, productivity as measured by the LFS was flat in 2025, but when measured using HMRC data it increased by 2%, in line with what we see in the US and what the OBR and the Bank of England say the UK seems incapable of achieving. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/roundup-of-the-week-20-february-2026) > …expect lower interest rates, better growth, lower inflation, better fiscal data and maybe — just maybe — a mea culpa from the establishment admitting they got it all wrong. (Sorry, I was getting carried away. That last bit will never happen.) > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/roundup-of-the-week-20-february-2026) [**13 March**](https://www.noisecancelling.co/read/roundup-of-the-week-13-march-2026) **and** [**20 March 2026 · Roundups of the week**](https://www.noisecancelling.co/read/roundup-of-the-week-20-march-2026) — the contradiction, week after week, while the MPC held at 3.75%. > The labour market data once again showed a glaring contradiction between the LFS and HMRC series, with the former showing more people in work and the latter pretty much showing no change. > > — [Neil Woodford, Noise Cancelling](https://www.noisecancelling.co/read/roundup-of-the-week-20-march-2026) [**22 July 2026 · Why I can see UK interest rates below 3% next year**](https://www.noisecancelling.co/read/why-i-can-see-uk-interest-rates-below-3-next-year) — by July, the argument had reached its destination: the case for materially lower interest rates, built on the truer labour market data. And then, on 18 August, the ONS fessed up. _Earlier this week:_ [Fessing up: the UK is more productive than they thought](https://www.noisecancelling.co/read/fessing-up-uk-productivity) — 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. Somewhere, my favourite economist is smiling. The mea culpa I joked would never happen is, more or less, in print. Now the policy needs to catch up with the data.