Four-quarter US productivity growth above 5%

Where things stand

• Latest data. The latest BLS Productivity and Costs release (Q2 2026 revised, Sept 3, 2026) shows four-quarter nonfarm business productivity growth (Table 2) of 2.2% in Q2 2026. The prior quarters were 2.9% (Q1 2026), 2.5% (Q4 2025) and 2.5% (Q3 2025). The index for Q2 2026 is 120.0 (2017=100). The next release is due Nov 5, 2026 (https://www.bls.gov/news.release/prod2.htm).
• Near-term quarters can't plausibly cross. For Q3 2026, the four-quarter figure would need roughly 15–17% annualized growth in a single quarter. Q4 2026 would need about 8–9% annualized for two quarters, and Q1 2027 about 6% for three.
• Current conditions.
  • Q3 GDP is tracking about 3.4–3.7% (GDPNow and Goldman Sachs, early October).
  • Payroll growth is weak: +29K in September, unemployment 4.2%.
  • BEA's Sept 30 annual update raised Q1 and Q2 2026 GDP growth to 2.5% and 2.2%.
  • My rough estimate is that four-quarter productivity for Q3 2026 comes in around 2–2.5%.
• Trend. Productivity has averaged 2.1% a year since Q4 2019 and about 2.5% over the last three years.

Reference class (computed)

Revised history. I downloaded FRED's OPHNFB index (1947–2026Q2, https://fred.stlouisfed.org/graph/fredgraph.csv?id=OPHNFB) and computed four-quarter changes rounded to one decimal.

• 21 of 314 quarters exceed 5.0%.
• Since 1974, the only ones are 2002Q1, 2003Q4, 2009Q4–2010Q1 and 2020Q2–2021Q1. All of these came in recessions or the recoveries after them.

Real-time readings. I also pulled monthly ALFRED vintages from 1985 to 2026 (alfred.stlouisfed.org/graph/alfredgraph.csv?id=OPHNFB&vintage_date=…).

• Some crossings showed up only in real time: 2000Q2 was first printed at about 5.2% in Aug 2000, then revised to about 4%.
• Others were published soon after the quarter: 2002Q3 (Nov 2002), 2003Q4 (Feb 2004) and 2009Q4 (Feb 2010, 5.1%, confirmed at https://www.bls.gov/news.release/archives/prod2_02042010.pdf).
• The 2020 spike crossed only after revisions, in the Nov 2022 vintage.
• Revisions are large: the first print minus the latest value has a standard deviation of about 1.1–1.25 points.

Window base rate. Of the 21-quarter windows starting in an expansion between 1985 and 2021, 44% contained a reading above 5.0% that was published before the window's publication cutoff (102 overlapping windows, effectively about four episodes). Across all start dates the figure is 50%.

Simulation check. I took historical deviations from a centered 41-quarter mean of growth and added them to a 2.0–2.3% trend. Windows starting in an expansion between 1983 and 2021 then give 43–59% in revised data. Almost every recession that peaks in the first 16 quarters produces a crossing in revised data.

Paths and model

Main path to yes: a recession followed by the usual productivity rebound.

• Recession odds. The current setting is late-cycle: the expansion is 6.5 years old, the Fed hiked on Sept 16, 2026, the 10-year yield is above 5%, oil is near $100, and there is AI-capex bubble risk. Deloitte's Sept 29 downside scenario has a recession in 2027–28. I put the chance of a business-cycle peak by mid-2030 at about 48–54%.
• Conditional chance of a published crossing: about 0.7–0.75. The 2001, 2008 and 2020 recessions all produced one. The 1973–75, 1980–82 and 1990–91 recessions did not in real time, but that was at a lower productivity trend.
• Lag from peak to publication: roughly 7–12 quarters in recent episodes.

Second path: no recession, but a crossing anyway. This would come from AI-driven acceleration or measurement noise, as in 2000Q2. I put it at about 10–15% over the window. Metaculus's Labor Hub community expects +18.6% productivity by 2030 versus 2025 (July 2026 update); I treat that as an upside tail, not my central case.

Model output. A quarterly hazard model (recession hazard, a lag distribution, and a rising no-recession hazard) gives about 38–52% by end-2031, centred near 44%. The series stays near zero through 2027 and then rises about 3–4 points per quarter from 2029.

Strongest case against

• Hiring is currently low-hire, low-fire, and low immigration makes labor scarce. Both point to firms holding on to staff in a downturn, which would mute the productivity spike.
• A recession that starts after about 2029, or one whose spike only appears after revisions (as in 2020), would publish too late.
• Outside recoveries, readings above 5% have been very rare since 1974.

I shaded my estimate below the simulation's 50–60% for these reasons.

Other sources

• CNBC, Oct 2, 2026 (September jobs report).
• Penn Mutual, Sept 21, 2026 (Fed hike).
• BEA, Sept 30, 2026: https://www.bea.gov/news/2026/gdp-third-estimate-industries-corporate-profits-state-gdp-and-state-personal-income-2nd
• Goldman Sachs Asset Management, Market Pulse, Oct 2026.
• Metaculus Labor Hub, July 2026 update.

Question. By the last day of each quarter, has BLS reported four-quarter nonfarm business productivity growth above 5% for any quarter?

Definitions.

• Productivity: BLS's seasonally adjusted U.S. nonfarm business sector labor productivity (real output per hour worked by all workers), not total factor productivity.
• Four-quarter growth: the percent change from the corresponding quarter of the previous year, as reported in the labor productivity column of Table 2 of Productivity and Costs. This is the change from quarter t−4 to quarter t, not the annualized change from the preceding quarter or the change between annual averages.
• Above 5%: strictly greater than 5.0% at BLS's published precision. A reported 5.0% does not qualify; do not infer a crossing from unrounded or independently calculated values.

Resolves YES if an official BLS publication reports a qualifying rate for a quarter ending after the opening date, under the Observation period convention. Preliminary estimates, revised estimates, and historical revisions all count for eligible quarters. No attribution to AI is required. A subsequent downward revision does not undo YES under the Resolution window convention.

Resolves NO if no qualifying report has been published by the horizon. Exclude forecasts, annual-average growth, annualized one-quarter growth, other sectors, and revisions or republications solely concerning quarters ending on or before the opening date.

Resolving source. BLS only: the Productivity and Costs release, Table 2, section “Percent change from corresponding quarter of previous year,” or the same measure in its official Major Sector Labor Productivity data. A successor BLS table counts only if it reports the same measure. Use dated BLS releases, archived releases, or documented BLS data updates to establish publication timing. The named-source restriction is narrower than Public confirmation (A2); apply Resolution date (A3), the Resolution window, and the Observation period conventions.