thesisPrototype

Government data

US JOLTS job openings, September 2026

What will BLS first report for total nonfarm job openings, seasonally adjusted, for September 2026 in the Job Openings and Labor Turnover Survey release?

Forecast

7.3M
6.7M–8M
80% prediction interval
gpt-5.5Run record ↗
6.36.97.68.2December 2025July 2026Nov 20267.3M
historyforecast path80% interval
Probability distribution

Cumulative probability

Chance that the outcome is at or below each value.

0%25%50%75%100%5.7M7.3M8.9MForecast value

Shaded band: 80% interval (6.7M–8M). Dashed line: point estimate (7.3M).

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

Analysis

§September 2026 BLS JOLTS job openings forecast

The resolver is the BLS JOLTS first-print preliminary total nonfarm job openings level, seasonally adjusted, for September 2026. The BLS schedule page lists September 2026 JOLTS for release on November 3, 2026 at 10:00 AM ET, which is inside the registered expected release window of 2026-11-01 to 2026-11-09, so resolutionDate is 2026-11-03.

Variant discipline: all anchors use the seasonally adjusted total nonfarm job openings level from BLS JOLTS Table A or equivalent first-print release table. The printed unit is thousands; this cell reports millions using the registered 0.001 transform.

▸ Reported tool use: official.lookupmodel report
BLS JOLTS release calendar lookup for September 2026
↳ Fetched schedule rows: July 2026 release Sep. 01, 2026 at 10:00 AM; August 2026 release Sep. 29, 2026 at 10:00 AM; September 2026 release Nov. 03, 2026 at 10:00 AM.
▸ Reported tool use: official.lookupmodel report
BLS archived JOLTS Table A first-print total job openings for late 2025 and early 2026
↳ Fetched first-print total job openings, seasonally adjusted, in thousands: December 2025 6542, January 2026 6946, February 2026 6882, March 2026 6866.
▸ Reported tool use: official.lookupmodel report
BLS archived JOLTS Table A first-print total job openings for April through July 2026
↳ Fetched first-print total job openings, seasonally adjusted, in thousands: April 2026 7618, May 2026 7594, June 2026 7359, July 2026 7271.
▸ Reported tool use: official.lookupmodel report
BLS Employment Situation August 2026 labor-market cross-check
↳ Fetched August 2026 labor-market values: total nonfarm payroll employment increased 162000, unemployment rate was 4.1 percent, June payroll change was revised to +31000, and July payroll change was revised to +21000.

The reference class and base rate are recent same-series first prints: December 2025 through July 2026 total openings ran from 6.542 million to 7.618 million, with the latest July print at 7.271 million. A random-walk persistence prior around 7.27 million is the chosen time-series model because JOLTS is a level series with strong local persistence.

Prior/update/interval: persistence prior model is a random walk anchored on the July 2026 first print of 7.271 million; historical sample is BLS first-print monthly openings from December 2025 through July 2026; adjustment components are +0.03 million for the August payroll rebound and stable 4.1 percent unemployment, +0.02 million for current openings still above early-2026 levels, and roughly 0.00 million for policy mechanism effects because no direct September hiring mandate applies. Successive changes were +0.404, -0.064, -0.016, +0.752, -0.024, -0.235, -0.088 million; one-month sample stdev is 0.347, so two-month horizon sigma = sqrt(2)*0.347 = 0.490. The 80 percent half-width is 1.28*sigma = 1.28*0.490 = 0.627 million. Point is 7.271 + 0.05 = 7.32; interval is 7.32 +/- 0.63 = [6.69, 7.95].

Upside risk would come from August and September openings rising alongside the 162000 August payroll gain, especially in food services, education, construction, or health care; a September first print near 8.0 million or above would land above the interval. Downside risk is renewed hiring freezes or a pullback in professional and business services; a broad demand drop below about 6.7 million would land outside the interval on the low side.

Review disposition: accepted both optional reviewer suggestions by explicitly tying the November 3, 2026 release date to the registered expected release window and naming the random-walk persistence prior as the chosen time-series model; no required fixes were raised.

Key drivers

  • July first-print openings were 7.271 million, close to the recent 7.3 million center
  • Recent first-print changes are volatile, with a 0.752 million April jump followed by smaller declines
  • August payroll growth of 162000 and 4.1 percent unemployment reduce near-term downside pressure
  • Low-hire, low-fire conditions argue against a sharp September breakout

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 Job Openings and Labor Turnover Survey first-print release
Resolution date
November 3, 2026· outcome not recorded
Resolution rule
Resolve to the first-published preliminary BLS Job Openings and Labor Turnover Survey total nonfarm job openings level, seasonally adjusted, for September 2026, from Table A or Table 1 of the initial release. The printed level is in thousands and is multiplied by 0.001 to millions. Ignore later monthly revisions, annual benchmark revisions, corrections after the first official print, and database values revised after the release.
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 publishable with only minor clarity issues; resolver, unit, first-print rule, prior/update math, interval, tails, and JSON fields are coherent.

  • info optional_suggestion: State explicitly that the 2026-11-03 release date is inside the registered 2026-11-01 to 2026-11-09 expected release window.
  • info optional_suggestion: Clarify that the persistence/random-walk prior is the chosen time-series model rather than leaving 'model prior' implicit.

disposition not applicable: Review disposition: accepted both optional reviewer suggestions by explicitly tying the November 3, 2026 release date to the registered expected release window and naming the random-walk persistence prior as the chosen time-series model; no required fixes were raised.

disposition not applicable: Review disposition: accepted both optional reviewer suggestions by explicitly tying the November 3, 2026 release date to the registered expected release window and naming the random-walk persistence prior as the chosen time-series model; no required fixes were raised.

Activity artifacts

Complete original trace

§September 2026 BLS JOLTS job openings forecast

The resolver is the BLS JOLTS first-print preliminary total nonfarm job openings level, seasonally adjusted, for September 2026. The BLS schedule page lists September 2026 JOLTS for release on November 3, 2026 at 10:00 AM ET, which is inside the registered expected release window of 2026-11-01 to 2026-11-09, so resolutionDate is 2026-11-03.

Variant discipline: all anchors use the seasonally adjusted total nonfarm job openings level from BLS JOLTS Table A or equivalent first-print release table. The printed unit is thousands; this cell reports millions using the registered 0.001 transform.

▸ Reported tool use: official.lookupmodel report
BLS JOLTS release calendar lookup for September 2026
↳ Fetched schedule rows: July 2026 release Sep. 01, 2026 at 10:00 AM; August 2026 release Sep. 29, 2026 at 10:00 AM; September 2026 release Nov. 03, 2026 at 10:00 AM.
▸ Reported tool use: official.lookupmodel report
BLS archived JOLTS Table A first-print total job openings for late 2025 and early 2026
↳ Fetched first-print total job openings, seasonally adjusted, in thousands: December 2025 6542, January 2026 6946, February 2026 6882, March 2026 6866.
▸ Reported tool use: official.lookupmodel report
BLS archived JOLTS Table A first-print total job openings for April through July 2026
↳ Fetched first-print total job openings, seasonally adjusted, in thousands: April 2026 7618, May 2026 7594, June 2026 7359, July 2026 7271.
▸ Reported tool use: official.lookupmodel report
BLS Employment Situation August 2026 labor-market cross-check
↳ Fetched August 2026 labor-market values: total nonfarm payroll employment increased 162000, unemployment rate was 4.1 percent, June payroll change was revised to +31000, and July payroll change was revised to +21000.

The reference class and base rate are recent same-series first prints: December 2025 through July 2026 total openings ran from 6.542 million to 7.618 million, with the latest July print at 7.271 million. A random-walk persistence prior around 7.27 million is the chosen time-series model because JOLTS is a level series with strong local persistence.

Prior/update/interval: persistence prior model is a random walk anchored on the July 2026 first print of 7.271 million; historical sample is BLS first-print monthly openings from December 2025 through July 2026; adjustment components are +0.03 million for the August payroll rebound and stable 4.1 percent unemployment, +0.02 million for current openings still above early-2026 levels, and roughly 0.00 million for policy mechanism effects because no direct September hiring mandate applies. Successive changes were +0.404, -0.064, -0.016, +0.752, -0.024, -0.235, -0.088 million; one-month sample stdev is 0.347, so two-month horizon sigma = sqrt(2)*0.347 = 0.490. The 80 percent half-width is 1.28*sigma = 1.28*0.490 = 0.627 million. Point is 7.271 + 0.05 = 7.32; interval is 7.32 +/- 0.63 = [6.69, 7.95].

Upside risk would come from August and September openings rising alongside the 162000 August payroll gain, especially in food services, education, construction, or health care; a September first print near 8.0 million or above would land above the interval. Downside risk is renewed hiring freezes or a pullback in professional and business services; a broad demand drop below about 6.7 million would land outside the interval on the low side.

Review disposition: accepted both optional reviewer suggestions by explicitly tying the November 3, 2026 release date to the registered expected release window and naming the random-walk persistence prior as the chosen time-series model; no required fixes were raised.

calibrated forecast · 80% CI
7.3M[6.7M · 8M]
Target metadata

Data point: bls.jolts.job_openings.september_2026.first_print

More government data forecasts