thesisPrototype

Government data

Australia Employment Change, September 2026

What will the Australian Bureau of Statistics first print for Labour Force, Australia series LF/M3.3.1599.20.AUS.M, seasonally adjusted employed persons month-over-month change, Australia, for September 2026, in thousands?

Forecast

18k
-29k–65k
80% prediction interval
gpt-5.5Run record ↗
-58-5.747992025-062026-07Oct 202618k
historyforecast path80% interval
Probability distribution

Cumulative probability

Chance that the outcome is at or below each value.

0%25%50%75%100%-99k18k135kForecast value

Shaded band: 80% interval (-29k–65k). Dashed line: point estimate (18k).

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

Analysis

§Australia September 2026 employment change first-print forecast

Framing and exact resolver: this is the ABS Labour Force, Australia first print for September 2026, series LF/M3.3.1599.20.AUS.M, seasonally adjusted employed persons, Australia, month-over-month change in thousands. Anchors use the same seasonally adjusted employed-person variant, not trend, original, hours, unemployment, or participation.

▸ Reported tool use: official.lookupmodel report
Inspect ABS Labour Force, Australia current release page future-release section and Labour Force Modernisation schedule article.
↳ Fetched official schedule numbers: Labour Force, Australia, August 2026 release date 24/09/2026 11:30am AEST; September 2026 release date 15/10/2026 11:30am AEDT; October 2026 release date 19/11/2026 11:30am AEDT. The registered target contract lists resolutionDate 2026-10-28.
▸ Reported tool use: official.lookupmodel report
Inspect ABS Labour Force, Australia July 2026 release key statistics and employment table.
↳ Fetched latest official print numbers: June 2026 seasonally adjusted employed people 14,823,100; July 2026 seasonally adjusted employed people 14,807,200; July monthly change -15,800, i.e. -15.8 thousand; July unemployment rate 4.5 percent; July trend employment monthly change +29,300, i.e. +29.3 thousand.
▸ Reported tool use: official.lookupmodel report
Read ABS July 2026 release employment table for same-variant recent seasonally adjusted employed-person levels.
↳ Fetched recent seasonally adjusted levels in thousands: 2025-05 14594.6, 2025-06 14569.4, 2025-07 14615.4, 2025-08 14600.8, 2025-09 14610.1, 2025-10 14643.3, 2025-11 14604.1, 2025-12 14661.4, 2026-01 14697.5, 2026-02 14719.6, 2026-03 14739.6, 2026-04 14704.7, 2026-05 14742.8, 2026-06 14823.1, 2026-07 14807.2.
▸ Reported tool use: python.computemodel report
Compute month-over-month changes and dispersion from the fetched ABS seasonally adjusted levels, using the ABS headline July change where available.
↳ Computed same-variant employment changes in thousands: 2025-06 -25.2, 2025-07 +46.0, 2025-08 -14.6, 2025-09 +9.3, 2025-10 +33.2, 2025-11 -39.2, 2025-12 +57.3, 2026-01 +36.1, 2026-02 +22.1, 2026-03 +20.0, 2026-04 -34.9, 2026-05 +38.1, 2026-06 +80.3, 2026-07 -15.8; mean +15.19, median +21.05, sample sigma 36.48.

Base rate / reference class: the recent official-source reference class is 14 same-series seasonally adjusted monthly employment changes from 2025-06 through 2026-07. The base rate is a positive but noisy monthly gain: mean +15.2 thousand, median +21.1 thousand, with swings from -39.2 to +80.3 thousand.

Level, momentum, one-off, and policy-mechanism effects: the level remains high at 14,807.2 thousand employed people in July, while momentum is mixed because the +80.3 thousand June surge was followed by -15.8 thousand in July. One-off survey and rotation effects are material in this series, especially around large adjacent moves. No policy mechanism suggests a discrete September employment break, and the July unemployment rate at 4.5 percent argues against a large positive demand surprise.

Prior/update/interval: prior = persistence toward the recent official-source mean/median reference class, with historical sample 14 monthly changes from 2025-06 through 2026-07. Adjustment components = +15.2 thousand base-rate mean, +4.8 thousand for the +21.1 thousand median and +29.3 thousand July trend gain, -2.0 thousand for the July negative SA print and 4.5 percent unemployment, giving point = 15.2 + 4.8 - 2.0 = 18.0 thousand. Interval method uses the values themselves for this change/flow series; sigma = 36.48 thousand, and 80% half-width = 1.28*sigma = 1.28*36.48 = 46.69 thousand. Final implied bounds = 18.0 - 46.69 = -28.69 and 18.0 + 46.69 = 64.69, rounded to [-28.7, 64.7].

Counter-considerations: upside risk is another broad hiring rebound or favourable survey rotation-group draw after July, which would land above the interval if September employment change exceeds +64.7 thousand. Downside risk is a sharper hiring slowdown or payback from June and July volatility, which would land below the interval if the first print is less than -28.7 thousand. An outside the interval outcome would most likely require a survey shock comparable to the largest recent positive or negative months.

Prior-run update: the public August target run used a recent-mean strategy through June 2026 and centered near +21 thousand; adding the official July -15.8 thousand print and the current ABS release table lowers the center modestly to +18.0 thousand and leaves realized dispersion near the prior run's broad interval scale.

Review disposition: accepted the resolver critique by narrowing the resolution rule to the registered ledger contract date rather than embedding the observed schedule discrepancy as a rule exception. Accepted the optional sample-label and mechanism wording improvements; forecast numbers are unchanged.

Key drivers

  • Recent same-series mean remains positive but volatile
  • July first print was negative after a strong June gain
  • Trend employment still rose in July
  • No August 2026 official print was available at run time
  • Cell follows registered ledger resolution date despite observed schedule evidence

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
Australian Bureau of Statistics Data API, Labour Force LF/M3.3.1599.20.AUS.M
Resolution date
October 28, 2026· outcome not recorded
Resolution rule
Resolve to the first official ABS observation for September 2026 in Labour Force series LF/M3.3.1599.20.AUS.M: seasonally adjusted employed persons, Australia, month-over-month change, in thousands. Use the first ABS Labour Force, Australia release/API print for the September 2026 reference period; ignore later revisions or corrections. This cell keeps the registered target resolutionDate of 2026-10-28 under the ledger target contract.
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 · fast · v2.5.11

pre-submit review · completed

Draft is publication-ready on forecast substance, but the resolver date discrepancy needs a blocking fix because it conflicts with the canonical ledger contract.

  • blocking resolver: The resolutionRule says the official ABS page currently schedules the September 2026 first print for 15 October 2026 while the registered contract sets resolutionDate to 2026-10-28; this creates an unresolved contract discrepancy inside the cell.
  • info optional_suggestion: The prior/update/interval step is strong; consider making the historical sample label explicitly say 14 monthly changes from 2025-06 through 2026-07.
  • info optional_suggestion: The tail scenarios are adequate, but could name concrete labor-market mechanisms such as survey rotation, hiring slowdown, or rebound from prior-month volatility.

disposition accepted: Review disposition: accepted the resolver critique by narrowing the resolution rule to the registered ledger contract date rather than embedding the observed schedule discrepancy as a rule exception. Accepted the optional sample-label and mechanism wording improvements; forecast numbers are unchanged.

disposition not applicable: Review disposition: accepted the resolver critique by narrowing the resolution rule to the registered ledger contract date rather than embedding the observed schedule discrepancy as a rule exception. Accepted the optional sample-label and mechanism wording improvements; forecast numbers are unchanged.

disposition not applicable: Review disposition: accepted the resolver critique by narrowing the resolution rule to the registered ledger contract date rather than embedding the observed schedule discrepancy as a rule exception. Accepted the optional sample-label and mechanism wording improvements; forecast numbers are unchanged.

Activity artifacts

Complete original trace

§Australia September 2026 employment change first-print forecast

Framing and exact resolver: this is the ABS Labour Force, Australia first print for September 2026, series LF/M3.3.1599.20.AUS.M, seasonally adjusted employed persons, Australia, month-over-month change in thousands. Anchors use the same seasonally adjusted employed-person variant, not trend, original, hours, unemployment, or participation.

▸ Reported tool use: official.lookupmodel report
Inspect ABS Labour Force, Australia current release page future-release section and Labour Force Modernisation schedule article.
↳ Fetched official schedule numbers: Labour Force, Australia, August 2026 release date 24/09/2026 11:30am AEST; September 2026 release date 15/10/2026 11:30am AEDT; October 2026 release date 19/11/2026 11:30am AEDT. The registered target contract lists resolutionDate 2026-10-28.
▸ Reported tool use: official.lookupmodel report
Inspect ABS Labour Force, Australia July 2026 release key statistics and employment table.
↳ Fetched latest official print numbers: June 2026 seasonally adjusted employed people 14,823,100; July 2026 seasonally adjusted employed people 14,807,200; July monthly change -15,800, i.e. -15.8 thousand; July unemployment rate 4.5 percent; July trend employment monthly change +29,300, i.e. +29.3 thousand.
▸ Reported tool use: official.lookupmodel report
Read ABS July 2026 release employment table for same-variant recent seasonally adjusted employed-person levels.
↳ Fetched recent seasonally adjusted levels in thousands: 2025-05 14594.6, 2025-06 14569.4, 2025-07 14615.4, 2025-08 14600.8, 2025-09 14610.1, 2025-10 14643.3, 2025-11 14604.1, 2025-12 14661.4, 2026-01 14697.5, 2026-02 14719.6, 2026-03 14739.6, 2026-04 14704.7, 2026-05 14742.8, 2026-06 14823.1, 2026-07 14807.2.
▸ Reported tool use: python.computemodel report
Compute month-over-month changes and dispersion from the fetched ABS seasonally adjusted levels, using the ABS headline July change where available.
↳ Computed same-variant employment changes in thousands: 2025-06 -25.2, 2025-07 +46.0, 2025-08 -14.6, 2025-09 +9.3, 2025-10 +33.2, 2025-11 -39.2, 2025-12 +57.3, 2026-01 +36.1, 2026-02 +22.1, 2026-03 +20.0, 2026-04 -34.9, 2026-05 +38.1, 2026-06 +80.3, 2026-07 -15.8; mean +15.19, median +21.05, sample sigma 36.48.

Base rate / reference class: the recent official-source reference class is 14 same-series seasonally adjusted monthly employment changes from 2025-06 through 2026-07. The base rate is a positive but noisy monthly gain: mean +15.2 thousand, median +21.1 thousand, with swings from -39.2 to +80.3 thousand.

Level, momentum, one-off, and policy-mechanism effects: the level remains high at 14,807.2 thousand employed people in July, while momentum is mixed because the +80.3 thousand June surge was followed by -15.8 thousand in July. One-off survey and rotation effects are material in this series, especially around large adjacent moves. No policy mechanism suggests a discrete September employment break, and the July unemployment rate at 4.5 percent argues against a large positive demand surprise.

Prior/update/interval: prior = persistence toward the recent official-source mean/median reference class, with historical sample 14 monthly changes from 2025-06 through 2026-07. Adjustment components = +15.2 thousand base-rate mean, +4.8 thousand for the +21.1 thousand median and +29.3 thousand July trend gain, -2.0 thousand for the July negative SA print and 4.5 percent unemployment, giving point = 15.2 + 4.8 - 2.0 = 18.0 thousand. Interval method uses the values themselves for this change/flow series; sigma = 36.48 thousand, and 80% half-width = 1.28*sigma = 1.28*36.48 = 46.69 thousand. Final implied bounds = 18.0 - 46.69 = -28.69 and 18.0 + 46.69 = 64.69, rounded to [-28.7, 64.7].

Counter-considerations: upside risk is another broad hiring rebound or favourable survey rotation-group draw after July, which would land above the interval if September employment change exceeds +64.7 thousand. Downside risk is a sharper hiring slowdown or payback from June and July volatility, which would land below the interval if the first print is less than -28.7 thousand. An outside the interval outcome would most likely require a survey shock comparable to the largest recent positive or negative months.

Prior-run update: the public August target run used a recent-mean strategy through June 2026 and centered near +21 thousand; adding the official July -15.8 thousand print and the current ABS release table lowers the center modestly to +18.0 thousand and leaves realized dispersion near the prior run's broad interval scale.

Review disposition: accepted the resolver critique by narrowing the resolution rule to the registered ledger contract date rather than embedding the observed schedule discrepancy as a rule exception. Accepted the optional sample-label and mechanism wording improvements; forecast numbers are unchanged.

calibrated forecast · 80% CI
18k[-29k · 65k]
Target metadata

Data point: abs.labour.employment_change.australia.september_2026.first_print

More government data forecasts