Four-quarter US productivity growth above 10%
Will BLS report four-quarter nonfarm business productivity growth above 10% for any quarter?
Chance by Q4 2031
5%
Cumulative probability
- Latest forecast
| Year | Q1 | Q2 | Q3 | Q4 |
|---|---|---|---|---|
| 2026 | <1% | |||
| 2027 | <1% | <1% | <1% | <1% |
| 2028 | <1% | <1% | <1% | <1% |
| 2029 | <1% | 1% | 1% | 2% |
| 2030 | 2% | 2% | 3% | 3% |
| 2031 | 3% | 4% | 4% | 5% |
Model reasoning
Since records began in 1948, official American figures have never shown output per hour of work rising more than about 7% in a year, and the latest reading is about 2%. Topping 10% by 2031 would almost certainly need an unprecedented surge in automation driven by artificial intelligence, which most economists consider unlikely this soon. What would most raise this forecast is a rapid wave of office-job automation: working hours falling while output keeps growing strongly.
Where things stand
- Latest BLS reading. The newest Productivity and Costs release (Table 2, revised Q2 2026, dated Sept 3, 2026) shows nonfarm business labor productivity up 2.2% from a year earlier in Q2 2026. Q1 2026 was 2.9% and 2025 as a whole was 2.1% (Table 2).
- Release schedule. The Q3 2026 preliminary estimate comes out Nov 5, 2026 and the revision on Dec 8. After that BLS publishes on its usual quarterly cycle (schedule). So by the 2031-12-31 horizon, data through Q3 2031 will count: 21 eligible quarters.
- The near quarters are effectively ruled out. I computed this from the index levels in the historical file:
- For Q3 2026 to exceed 10%, the index would have to reach 130.8 or more against 118.9 a year earlier. That is a 9.0% jump in one quarter from Q2 2026's 120.0, about 41% annualized.
- Q4 2026 would need about 9.4% growth over two quarters.
- Meanwhile the Atlanta Fed's GDPNow puts Q3 2026 growth at about 3.6%, and September payrolls rose only 29k. That points to solid but ordinary productivity, roughly 3–4% annualized.
Base rate
I downloaded the BLS historical file labor-productivity-major-sectors.xlsx (MachineReadable tab; nonfarm business, all workers, "% Change same quarter 1 year ago"). It covers 314 quarters from 1948Q1 to 2026Q2:
- mean 2.16%, standard deviation 1.82 points;
- maximum 7.2% (1950Q4), next 7.1% (2020Q3, at the height of the pandemic shift in which jobs were lost disproportionately in low-productivity work);
- only 2 quarters ever above 7%, and none above 8%.
A reading above 10% would be about 4.3 standard deviations above the mean and 40% above the post-war record. On historical patterns alone (recessions, pandemics, data artifacts), I put the chance in five years at roughly 0.3–0.5%.
The main path to YES: an AI economic takeoff
- How it would happen. Output would have to jump while hours stay flat or fall sharply, for example through mass automation of office work.
- Anthropic's scenario model. Korinek et al. (Anthropic Institute, Sept 2026) set out three scenarios (scenario explorer, paper):
- "Extreme": 15.4% GDP growth in the 12 months to 2030, with office-type ("cognitive") employment down 21.5%. Productivity growth would be far above 10%.
- "Substantial": about 5.4% GDP growth, which would not qualify.
- The authors attach no probabilities. About 10% of surveyed American adults hold views consistent with the extreme case.
- Signs that AI capabilities are moving fast:
- Dario Amodei's "We Must Pace the Frontier" (Sept 12, 2026) says recursive self-improvement began around summer 2026.
- GPT-6 Astra was launched Sept 3, 2026 and promoted as the start of an "AGI era".
- The government restricted the releases of Mythos 5 and GPT-5.6.
- Metaculus's AGI estimate is now around Jan 2031, with a 25th percentile of about April 2028.
- Reasons to doubt a fast economic effect:
- Measured productivity is still about 2%.
- In the Forecasting Research Institute's survey (NBER w35046, Apr 2026), economists' median labor productivity growth for 2030 is 2.0%, or 3.2% even in their "rapid AI" scenario. Their 90th percentile for GDP growth under rapid AI is about 6.4%. AI-industry experts and superforecasters have fatter tails, with some weight above 10%.
- The Metaculus community expects only about +15–19% cumulative productivity growth for 2025–2030, roughly 3% a year.
- Deloitte (Sept 29, 2026) projects about 2% GDP growth through 2031 and sees little AI productivity evidence so far.
- The labs themselves are calling for a slowdown, and adoption, compute, regulation and the physical economy all limit how fast automation can spread.
Estimate
I put 25% on AGI-level systems by early 2028. Given that, I give about a 15% chance that a 10%+ year shows up in BLS data by Q3 2031. Adding later-AGI paths, an AI-driven layoff spike, and non-AI shocks brings the total to about 4.7% by end-2031.
I built the series from quarterly hazards: essentially zero through 2027, about 0.1–0.16% per quarter in 2028, and rising to about 0.48% per quarter in 2031. Cumulative values are about 1.6% by end-2029 and 3.0% by end-2030.
Strongest case against: if recursive self-improvement really is under way, an extreme-type scenario by 2029–2030 could deserve 10%+. Conversely, a 10% reading has no precedent even in the pandemic, and economists put it near zero. I stay between the two, leaning toward the historical record because measured data still show no acceleration.
Resolution criteria
Question. By the last day of each quarter, has BLS reported four-quarter U.S. nonfarm business productivity growth above 10% for any quarter?
Definitions.
- Productivity: BLS nonfarm business sector labor productivity, measured as real output per hour worked by all workers, seasonally adjusted; not total factor productivity.
- Four-quarter growth: the percentage change from the corresponding quarter of the previous year, as reported in the “Labor productivity” column of the “Percent change from corresponding quarter of previous year” panel in Table 2 of Productivity and Costs. This is not the annualized change from the immediately preceding quarter or the change between annual averages.
- Above 10%: strictly greater than 10.0% at BLS's published precision for the growth rate. A reported 10.0% does not qualify; do not infer a qualifying rate from rounded index levels.
Resolves YES if a BLS publication within the Resolution window reports a qualifying rate for a quarter ending after the opening date, pursuant to the Observation period convention. Preliminary estimates, regular revisions, and benchmark revisions count. No attribution to AI is required. A qualifying published estimate remains sufficient even if later revised below the threshold.
Resolves NO if no qualifying report has been published by the horizon. Exclude figures for quarters ending before the opening date, forecasts, annual-average growth, one-quarter annualized growth, and figures for the business sector including farms, manufacturing, nonfinancial corporations, or individual industries rather than the aggregate nonfarm business sector.
Resolving source. U.S. Bureau of Labor Statistics only: Productivity and Costs, Table 2: Nonfarm business sector, its dated release archives, or the corresponding BLS historical data file Labor productivity and costs, major sectors. A successor BLS table or publication counts only if it reports the same sector, measure, and four-quarter comparison. Apply the Resolution date convention (A3) to the first qualifying BLS publication, and the Resolution window and Observation period conventions. Secondary reporting alone cannot resolve the question.