thesisPrototype

Government data

US Nonfarm Productivity Q3 2026 Preliminary

What will BLS first report for nonfarm business sector labor productivity, seasonally adjusted percent change from previous quarter at annual rate, in 2026 Q3 preliminary Table 2?

Forecast

+1.7%
-1.7%–+5.1%
80% prediction interval
gpt-5.5Run record ↗

Cumulative probability

Chance that the outcome is at or below each value.

0%25%50%75%100%-6.8%+1.7%+10.2%Forecast value

Shaded band: 80% interval (-1.7%–+5.1%). Dashed line: point estimate (+1.7%).

Derived from the point estimate and 80% interval; the agent did not report a full distribution.

Analysis

§US nonfarm business labor productivity, 2026 Q3 preliminary first print

Framing and exact resolver: target is BLS Productivity and Costs Table 2, nonfarm business sector Labor productivity, seasonally adjusted percent change from previous quarter at an annual rate, for 2026 Q3 preliminary. The FRED/BLS series code used for history is PRS85006092; resolution remains the BLS first-print Table 2, not FRED.

▸ Reported tool use: official.lookupmodel report
Checked BLS Productivity and Costs release schedule for the Third Quarter 2026 preliminary release.
↳ BLS schedule lists Third Quarter 2026 (P) on Nov. 05, 2026 at 08:30 AM and Third Quarter 2026 (R) on Dec. 08, 2026 at 08:30 AM.
▸ Reported tool use: official.lookupmodel report
Fetched current BLS Productivity and Costs Table 2 for nonfarm business sector labor productivity and components.
↳ BLS Table 2 shows 2026 Q1 labor productivity 0.3, output 1.0, hours worked 0.7; 2025 Q4 labor productivity 1.6, 2025 Q3 5.2, 2025 Q2 4.2, 2025 Q1 -0.9.
▸ Reported tool use: history.lookupmodel report
Fetched FRED mirror of BLS PRS85006092, Percent Change at Annual Rate, Seasonally Adjusted, quarterly.
↳ FRED PRS85006092 reports Q1 2026 0.3, Q4 2025 1.6, Q3 2025 5.2, Q2 2025 4.2, Q1 2025 -0.9; the date range is 1947-04-01 to 2026-01-01 and last updated 2026-06-04 7:41 AM CDT.
▸ Reported tool use: history.lookupmodel report
Read recent BLS/FRED history for the post-2020 reference class used to size uncertainty.
↳ For 2021Q1-2026Q1 PRS85006092 values are 2.9, 0.8, -2.4, 2.3, -4.9, -2.9, 0.2, 2.3, 0.9, 4.1, 5.1, 3.9, 0.2, 3.7, 3.7, 1.4, -0.9, 4.2, 5.2, 1.6, 0.3.

Base rate/reference class: the post-2020 nonfarm business productivity values average about 1.5 percent annualized, while the 2023-2025 subset is stronger at about 2.8 percent. I put more weight on the broader post-2020 base rate because quarterly productivity first prints are noisy and 2025 likely included tariff and inventory timing effects. The history is revised BLS/FRED public history used as a proxy for first-print volatility because a clean first-print vintage sample was not used.

Prior/update/interval: persistence/reference-class prior is the 2021Q1-2026Q1 PRS85006092 mean, 31.7/21 = 1.51, with no separate AR or econometric time-series model beyond this reference-class prior. I add +0.1 for mean reversion after the weak 2026 Q1 print and +0.1 for moderate trend productivity/capital deepening, giving point = 1.7. For this change-rate series I compute sigma from the fetched values themselves: squared deviations from 1.51 sum to about 146.4, variance = 146.4/21 = 6.97, sigma = 2.64. The 80 percent normal half-width is roughly 1.28*sigma = 1.28*2.64 = 3.38, so 1.7 +/- 3.4 gives -1.7 to 5.1 after one-decimal rounding.

Counter-considerations: upside risk is a Q3 output rebound with flat or falling hours, which would land above the interval if productivity prints above 5.1. Downside risk is weak real output with continued hours growth, which would land below the interval if productivity prints below -1.7. A large inventory/tariff timing swing or a recession-style hours adjustment is the main outside the interval scenario.

Review disposition: accepted the reviewer suggestions to clarify that revised BLS/FRED history is a proxy for first-print volatility, state that no separate AR or econometric model was used, and split the +0.2 adjustment into weak-Q1 mean reversion and moderate trend productivity components.

Key drivers

  • Recent revised productivity slowed to 0.3 percent in 2026 Q1 after strong 2025 readings
  • Post-2021 reference class remains volatile but centered near 1.5 percent annualized
  • Q3 first print will depend on early GDP/output and hours estimates available to BLS in November
  • AI/capital-deepening narrative supports a modest positive trend but not a large near-term jump

Tool evidence

This run has no captured tool responses. “Reported tool use” in the analysis is the model’s account, not an archived tool response.

Sources and resolution

Official source
U.S. Bureau of Labor Statistics Productivity and Costs news release, Table 2
Resolution date
November 5, 2026· outcome not recorded
Resolution rule
Resolve to the first BLS Productivity and Costs preliminary release for Third Quarter 2026, Table 2, Nonfarm business sector, row 2026 III, Labor productivity, seasonally adjusted percent change from previous quarter at annual rate. Use the one-decimal value printed in the first official release on 2026-11-05 and do not apply later revised releases.
Run details

The analysis is the model’s written report. Tool-use descriptions in that report are model claims; the activity artifacts contain the execution record.

thesis.analyst · gpt-5.5 · v2.2.0

pre-submit review · completed

Draft is publishable with no blocking issues; resolver, first-print rule, point, interval, and JSON fields are coherent, with only minor clarity improvements recommended.

  • info optional_suggestion: Clarify that the 2021Q1-2026Q1 history uses revised BLS/FRED values as a proxy for first-print volatility, or switch to first-print vintage history if available.
  • info optional_suggestion: State explicitly that no separate AR or econometric time-series model was used beyond the reference-class/persistence prior.
  • info optional_suggestion: Tighten the +0.2 adjustment by separating the weak-Q1 mean-reversion component from the AI/capital-deepening trend component.

disposition not applicable: Review disposition: accepted the reviewer suggestions to clarify that revised BLS/FRED history is a proxy for first-print volatility, state that no separate AR or econometric model was used, and split the +0.2 adjustment into weak-Q1 mean reversion and moderate trend productivity components.

disposition not applicable: Review disposition: accepted the reviewer suggestions to clarify that revised BLS/FRED history is a proxy for first-print volatility, state that no separate AR or econometric model was used, and split the +0.2 adjustment into weak-Q1 mean reversion and moderate trend productivity components.

disposition not applicable: Review disposition: accepted the reviewer suggestions to clarify that revised BLS/FRED history is a proxy for first-print volatility, state that no separate AR or econometric model was used, and split the +0.2 adjustment into weak-Q1 mean reversion and moderate trend productivity components.

Activity artifacts

Complete original trace

§US nonfarm business labor productivity, 2026 Q3 preliminary first print

Framing and exact resolver: target is BLS Productivity and Costs Table 2, nonfarm business sector Labor productivity, seasonally adjusted percent change from previous quarter at an annual rate, for 2026 Q3 preliminary. The FRED/BLS series code used for history is PRS85006092; resolution remains the BLS first-print Table 2, not FRED.

▸ Reported tool use: official.lookupmodel report
Checked BLS Productivity and Costs release schedule for the Third Quarter 2026 preliminary release.
↳ BLS schedule lists Third Quarter 2026 (P) on Nov. 05, 2026 at 08:30 AM and Third Quarter 2026 (R) on Dec. 08, 2026 at 08:30 AM.
▸ Reported tool use: official.lookupmodel report
Fetched current BLS Productivity and Costs Table 2 for nonfarm business sector labor productivity and components.
↳ BLS Table 2 shows 2026 Q1 labor productivity 0.3, output 1.0, hours worked 0.7; 2025 Q4 labor productivity 1.6, 2025 Q3 5.2, 2025 Q2 4.2, 2025 Q1 -0.9.
▸ Reported tool use: history.lookupmodel report
Fetched FRED mirror of BLS PRS85006092, Percent Change at Annual Rate, Seasonally Adjusted, quarterly.
↳ FRED PRS85006092 reports Q1 2026 0.3, Q4 2025 1.6, Q3 2025 5.2, Q2 2025 4.2, Q1 2025 -0.9; the date range is 1947-04-01 to 2026-01-01 and last updated 2026-06-04 7:41 AM CDT.
▸ Reported tool use: history.lookupmodel report
Read recent BLS/FRED history for the post-2020 reference class used to size uncertainty.
↳ For 2021Q1-2026Q1 PRS85006092 values are 2.9, 0.8, -2.4, 2.3, -4.9, -2.9, 0.2, 2.3, 0.9, 4.1, 5.1, 3.9, 0.2, 3.7, 3.7, 1.4, -0.9, 4.2, 5.2, 1.6, 0.3.

Base rate/reference class: the post-2020 nonfarm business productivity values average about 1.5 percent annualized, while the 2023-2025 subset is stronger at about 2.8 percent. I put more weight on the broader post-2020 base rate because quarterly productivity first prints are noisy and 2025 likely included tariff and inventory timing effects. The history is revised BLS/FRED public history used as a proxy for first-print volatility because a clean first-print vintage sample was not used.

Prior/update/interval: persistence/reference-class prior is the 2021Q1-2026Q1 PRS85006092 mean, 31.7/21 = 1.51, with no separate AR or econometric time-series model beyond this reference-class prior. I add +0.1 for mean reversion after the weak 2026 Q1 print and +0.1 for moderate trend productivity/capital deepening, giving point = 1.7. For this change-rate series I compute sigma from the fetched values themselves: squared deviations from 1.51 sum to about 146.4, variance = 146.4/21 = 6.97, sigma = 2.64. The 80 percent normal half-width is roughly 1.28*sigma = 1.28*2.64 = 3.38, so 1.7 +/- 3.4 gives -1.7 to 5.1 after one-decimal rounding.

Counter-considerations: upside risk is a Q3 output rebound with flat or falling hours, which would land above the interval if productivity prints above 5.1. Downside risk is weak real output with continued hours growth, which would land below the interval if productivity prints below -1.7. A large inventory/tariff timing swing or a recession-style hours adjustment is the main outside the interval scenario.

Review disposition: accepted the reviewer suggestions to clarify that revised BLS/FRED history is a proxy for first-print volatility, state that no separate AR or econometric model was used, and split the +0.2 adjustment into weak-Q1 mean reversion and moderate trend productivity components.

calibrated forecast · 80% CI
+1.7%[-1.7% · +5.1%]
Target metadata

Data point: bls.productivity.nonfarm_qoq_prelim.2026_q3.first_print

More government data forecasts