UK PRODUCTIVITY · Methods published 17 September; reference tables 6 October 2026
Counting working hours differently changes the size of the UK's measured productivity slowdown. In the Office for National Statistics' proposed component approach, average annual growth in output per hour for 2009–2019 is 1.3%, compared with 0.7% under the current approach. These are indicative methodological estimates, explicitly not official statistics, and do not represent new output produced in October.
The useful question is where the revision enters the calculation. Output per hour divides economic output by hours worked. Changing the denominator can alter the estimated productivity path even when the output series stays the same. This makes the new tables a way to investigate measurement, rather than a reason to announce an overnight improvement in the economy.
How much of the slowdown changes?
Average annual output-per-hour growth. Common scale: 0% to 2.5%.
1997–2007
2009–2019
| Approach | 1997–2007 | 2009–2019 | Later minus earlier |
|---|---|---|---|
| Current | 2.1% | 0.7% | −1.4 percentage points |
| Component | 2.0% | 1.3% | −0.7 percentage points |
The comparison uses two periods separated by the crisis year. Subtracting the two rates measures the slowdown in annual growth; it does not measure lost pounds of output. On these rounded inputs, its size is 0.7 percentage points smaller under the proposed method. That is a useful description of the methodological impact, with no implication that either rate is known to unlimited precision.
What is being measured differently?
ONS combines household surveys, employer surveys and administrative records to construct jobs, workers and hours. Explicit adjustments cover leave, sickness, holidays, overtime and furlough. The method does not change gross value added. Its intended whole-economy implementation is November 2026, with industry estimates in February 2027. Those dates describe the announced statistical rollout, not a forecast of economic performance. The methodology explains the proposed changes.
Why does a smaller hours denominator raise productivity?
Consider a purely illustrative economy producing 1,000 units of output. If measured hours are 100, output per hour is 10. If a better measurement finds 95 hours for the same output, the ratio is about 10.526. The measured level rises by approximately 5.26%, although output has not changed.
The formula is (100 ÷ 95 − 1) × 100. This invented example explains the arithmetic only. It is not an estimate of the UK's hours revision, and a change in one year's level alone does not establish a change in the growth rate.
Why keep output per worker and output per hour separate?
A worker is a person; a job is a position; an hour is an amount of work time. Someone with two jobs is still one worker. Two workforces with the same number of people can supply different hours. An hourly estimate can therefore move differently from a per-person estimate without either being mathematically inconsistent.
For a comparison dashboard, give each measure its own label and denominator. Do not replace an hourly series with a per-worker series mid-chart, or splice the proposed historical series onto the current series without marking the break.
What should a reader take away?
The economic reading depends on both activity and how it is measured. Here the strongest conclusion is that the estimated intensity of the productivity slowdown is sensitive to labor-input methods. The supplied comparison cannot establish that a particular technology, policy or industry caused the difference. A causal claim would require additional evidence.
For reproducible work, keep the methodology date, reference-table date, observation years and approach together. The 6 October release supplies absolute-value tables for the earlier methodology, with current-price and volume measures. Choose the volume measure for a real-productivity comparison, then retain the same approach across the full period.
Sources, dates and reuse
ONS, Component approach to measuring labour productivity, UK: 1997 to 2024 (17 September 2026); reference-table release notice (6 October 2026, 09:30 UK time); downloadable reference tables. Observations: 1997–2024; growth comparisons above: 1997–2007 and 2009–2019. Source check: 9 October 2026, 01:57 UTC. All comparisons and graphics are original; chart values are repeated in text or tables. Source: Office for National Statistics, licensed under the Open Government Licence v3.0. ONS reuse guidance permits commercial reuse with attribution, subject to exceptions. No agency logo, external photograph or third-party chart is reproduced; no official endorsement is implied.
