thesisPrototype

Government data

US continued claims, week ending 2026-09-19

What will the first-print U.S. Department of Labor Employment and Training Administration seasonally adjusted insured unemployment, continued claims series CCSA, be for week ending 2026-09-19, in millions?

Forecast

1.8M
1.7M–1.8M
80% prediction interval
gpt-5.5Run record ↗

Cumulative probability

Chance that the outcome is at or below each value.

0%25%50%75%100%1.7M1.8M1.9MForecast value

Shaded band: 80% interval (1.7M–1.8M). Dashed line: point estimate (1.8M).

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

Analysis

§Forecast CCSA for week ending 2026-09-19

Framing and exact resolver: this targets DOL ETA continued claims, also called seasonally adjusted insured unemployment, for week ending 2026-09-19. The resolving variant is CCSA, seasonally adjusted, advance first print, not the not-seasonally-adjusted count, insured unemployment rate, four-week average, current revised FRED value, or a later ALFRED vintage. The unit is the registered targetUnit millions.

▸ Reported tool use: local.target.lookupmodel report
Read the registered target file and generated ledger target entry for continued-claims-week-2026-09-19.
↳ Fetched target fields: catalogSlug continued-claims-week-2026-09-19; dataPointId dol.eta.continued_claims.sa.week_2026-09-19.first_print; unit millions; sourceSeriesId CCSA; transform factor 0.000001; expectedReleaseWindow start 2026-10-05 end 2026-10-09; registeredAtUtc 2026-09-14T20:23:15Z; targetContentHash 75a1aa15e93e29e83eb184f6b3af1b67c83c0bd5890038ce9f51827d8afba59e.
▸ Reported tool use: official.lookupmodel report
Opened the current DOL ETA Unemployment Insurance Weekly Claims News Release PDF dated 2026-09-10.
↳ The release reports advance seasonally adjusted insured unemployment for week ending 2026-08-29 at 1,774,000, down 1,000 from the revised 1,775,000 for 2026-08-22; the four-week moving average was 1,779,000 and the insured unemployment rate was 1.2 percent.
▸ Reported tool use: official.lookupmodel report
Read the DOL release's Seasonally Adjusted US Weekly UI Claims table for recent same-variant CCSA history.
↳ Recent CCSA values in thousands were 1,812 for 2026-06-13, 1,806 for 2026-06-20, 1,821 for 2026-06-27, 1,798 for 2026-07-04, 1,789 for 2026-07-11, 1,777 for 2026-07-18, 1,799 for 2026-07-25, 1,781 for 2026-08-01, 1,796 for 2026-08-08, 1,771 for 2026-08-15, 1,775 for 2026-08-22, and 1,774 for 2026-08-29.
▸ Reported tool use: official.lookupmodel report
Checked the DOL archive publication schedule and the FRED release calendar for UI Weekly Claims release timing.
↳ The DOL archive says the UI Weekly Claims News Release is published each Thursday at 8:30 AM Eastern and lists 1 non-Thursday 2026 exception, Wednesday 2026-11-25. The release calendar lists UI Weekly Claims Report releases on Thursday 2026-09-17, 2026-09-24, 2026-10-01, and 2026-10-08; the registered resolutionDate remains 2026-10-09.
▸ Reported tool use: official.lookupmodel report
Opened BLS current seasonal factors for weekly initial and continued claims.
↳ For continued claims, BLS lists seasonal factors of 0.945 for 2026-08-29, 0.912 for 2026-09-05, 0.903 for 2026-09-12, 0.881 for 2026-09-19, and 0.877 for 2026-09-26; for initial claims the latest DOL release reported 206,000 for 2026-09-05.

Reference class and base rate: the outside-view base rate is persistence in the same seasonally adjusted insured-unemployment series. The last 12 same-variant CCSA prints average 1.792 million, the median is 1.793 million, the latest four average 1.779 million, and the latest print is 1.774 million, so I anchor near the latest level rather than chasing the higher early-summer values.

Level, momentum, one-off, and policy-mechanism effects: level starts at 1.774 million. Momentum is flat to slightly negative, with the last four observations 1.796, 1.771, 1.775, and 1.774 million and a latest four-week average of 1.779 million. The stable 206,000 initial-claims print reduces near-term upside pressure. I found no federal extended-benefit or policy mechanism in the DOL release that should mechanically shift regular state-program seasonally adjusted insured unemployment.

Prior/update/interval: persistence prior = latest official CCSA first print of 1.774 million for week ending 2026-08-29; historical sample = fetched DOL same-variant weekly CCSA first-print levels from 2026-03-14 through 2026-08-29, using successive weekly changes. Adjustment components = +0.003 million toward the latest four-week average of 1.779, +0.002 million for possible post-Labor-Day reaccumulation, and +0.000 million policy-mechanism effect, giving point = 1.774 + 0.003 + 0.002 = 1.779 million. Successive-change dispersion over the fetched sample gives one-week sigma = 0.0195 million; for the three-week horizon to 2026-09-19, assuming independent weekly changes, sigma = sqrt(3)*0.0195 = 0.0337 million; 80% half-width = 1.28*sigma = 1.28*0.0337 = 0.043 million. Final implied bounds are 1.779 +/- 0.043 = [1.736, 1.822].

Counter-considerations: upside risk is a sequence of weak reemployment, delayed post-holiday filings, or longer benefit duration that would push CCSA above 1.822 million. Downside risk is sustained low initial claims and faster exits from insured unemployment, which would land below 1.736 million. Outside the interval would likely require a clear break from the recent 1.77 to 1.82 million range, not just ordinary one-week noise.

Update from prior run: the previous public Thesis continued-claims run for week ending 2026-09-12 used 1.779 million through 2026-08-22 and forecast 1.785 million. The new DOL 2026-09-10 print added week ending 2026-08-29 at 1.774 million and revised 2026-08-22 to 1.775 million, so I move the point down to 1.779 while keeping a similar volatility-based interval.

Review disposition: accepted the resolver critique by making the target-registration timing discrepancy explicit while preserving the registered slug, unit, dataPointId, source URL, and resolutionDate. Accepted the optional clarifications by naming the 2026-08-29 first-print sample endpoint and stating the independent weekly-change assumption behind the sqrt(3) horizon scaling.

Key drivers

  • Latest DOL first-print CCSA was 1.774 million for week ending 2026-08-29
  • The latest four-week average was 1.779 million, close to the latest level
  • Recent same-variant continued claims are range-bound around 1.77 to 1.82 million
  • Initial claims remain stable near 206 thousand, limiting near-term upside pressure
  • Target week has lower continued-claims seasonal factors, but the forecast is for the seasonally adjusted CCSA variant

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. Department of Labor Unemployment Insurance Weekly Claims release, preserved through ALFRED CCSA advance vintage
Resolution date
October 9, 2026· outcome not recorded
Resolution rule
Resolve from ALFRED graph CSV series CCSA for observation_date 2026-09-19, using the first ALFRED advance vintage value corresponding to the official U.S. Department of Labor ETA Unemployment Insurance Weekly Claims first print for seasonally adjusted insured unemployment. Transform the reported persons count by multiplying by 0.000001 to express it in millions and round to 0.001 million for display. Use the first print only and ignore later revisions or benchmark updates. Keep the registered resolutionDate 2026-10-09 and expectedReleaseWindow 2026-10-05 to 2026-10-09; the DOL Thursday cadence suggests the observation may first appear before that window, likely in the 2026-10-01 release, so this is recorded as a target-registration timing discrepancy rather than a change to the resolver.
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 largely publishable, but it should tighten the resolver/date discrepancy because the stated DOL cadence implies a first print before the registered resolution window.

  • warning resolver: The reasoning says the 2026-09-19 continued-claims first print may appear in the 2026-10-01 release, while the target contract uses resolutionDate 2026-10-09 and expectedReleaseWindow 2026-10-05 to 2026-10-09.
  • info optional_suggestion: The prior/update/interval step is strong; consider naming the fetched sample endpoint explicitly as 2026-08-29 first print to avoid any revised-vintage ambiguity.
  • info optional_suggestion: The interval method is coherent, but note that the sqrt(3) horizon assumes independent weekly changes.

disposition accepted: Review disposition: accepted the resolver critique by making the target-registration timing discrepancy explicit while preserving the registered slug, unit, dataPointId, source URL, and resolutionDate. Accepted the optional clarifications by naming the 2026-08-29 first-print sample endpoint and stating the independent weekly-change assumption behind the sqrt(3) horizon scaling.

disposition not applicable: Review disposition: accepted the resolver critique by making the target-registration timing discrepancy explicit while preserving the registered slug, unit, dataPointId, source URL, and resolutionDate. Accepted the optional clarifications by naming the 2026-08-29 first-print sample endpoint and stating the independent weekly-change assumption behind the sqrt(3) horizon scaling.

disposition not applicable: Review disposition: accepted the resolver critique by making the target-registration timing discrepancy explicit while preserving the registered slug, unit, dataPointId, source URL, and resolutionDate. Accepted the optional clarifications by naming the 2026-08-29 first-print sample endpoint and stating the independent weekly-change assumption behind the sqrt(3) horizon scaling.

Activity artifacts

Complete original trace

§Forecast CCSA for week ending 2026-09-19

Framing and exact resolver: this targets DOL ETA continued claims, also called seasonally adjusted insured unemployment, for week ending 2026-09-19. The resolving variant is CCSA, seasonally adjusted, advance first print, not the not-seasonally-adjusted count, insured unemployment rate, four-week average, current revised FRED value, or a later ALFRED vintage. The unit is the registered targetUnit millions.

▸ Reported tool use: local.target.lookupmodel report
Read the registered target file and generated ledger target entry for continued-claims-week-2026-09-19.
↳ Fetched target fields: catalogSlug continued-claims-week-2026-09-19; dataPointId dol.eta.continued_claims.sa.week_2026-09-19.first_print; unit millions; sourceSeriesId CCSA; transform factor 0.000001; expectedReleaseWindow start 2026-10-05 end 2026-10-09; registeredAtUtc 2026-09-14T20:23:15Z; targetContentHash 75a1aa15e93e29e83eb184f6b3af1b67c83c0bd5890038ce9f51827d8afba59e.
▸ Reported tool use: official.lookupmodel report
Opened the current DOL ETA Unemployment Insurance Weekly Claims News Release PDF dated 2026-09-10.
↳ The release reports advance seasonally adjusted insured unemployment for week ending 2026-08-29 at 1,774,000, down 1,000 from the revised 1,775,000 for 2026-08-22; the four-week moving average was 1,779,000 and the insured unemployment rate was 1.2 percent.
▸ Reported tool use: official.lookupmodel report
Read the DOL release's Seasonally Adjusted US Weekly UI Claims table for recent same-variant CCSA history.
↳ Recent CCSA values in thousands were 1,812 for 2026-06-13, 1,806 for 2026-06-20, 1,821 for 2026-06-27, 1,798 for 2026-07-04, 1,789 for 2026-07-11, 1,777 for 2026-07-18, 1,799 for 2026-07-25, 1,781 for 2026-08-01, 1,796 for 2026-08-08, 1,771 for 2026-08-15, 1,775 for 2026-08-22, and 1,774 for 2026-08-29.
▸ Reported tool use: official.lookupmodel report
Checked the DOL archive publication schedule and the FRED release calendar for UI Weekly Claims release timing.
↳ The DOL archive says the UI Weekly Claims News Release is published each Thursday at 8:30 AM Eastern and lists 1 non-Thursday 2026 exception, Wednesday 2026-11-25. The release calendar lists UI Weekly Claims Report releases on Thursday 2026-09-17, 2026-09-24, 2026-10-01, and 2026-10-08; the registered resolutionDate remains 2026-10-09.
▸ Reported tool use: official.lookupmodel report
Opened BLS current seasonal factors for weekly initial and continued claims.
↳ For continued claims, BLS lists seasonal factors of 0.945 for 2026-08-29, 0.912 for 2026-09-05, 0.903 for 2026-09-12, 0.881 for 2026-09-19, and 0.877 for 2026-09-26; for initial claims the latest DOL release reported 206,000 for 2026-09-05.

Reference class and base rate: the outside-view base rate is persistence in the same seasonally adjusted insured-unemployment series. The last 12 same-variant CCSA prints average 1.792 million, the median is 1.793 million, the latest four average 1.779 million, and the latest print is 1.774 million, so I anchor near the latest level rather than chasing the higher early-summer values.

Level, momentum, one-off, and policy-mechanism effects: level starts at 1.774 million. Momentum is flat to slightly negative, with the last four observations 1.796, 1.771, 1.775, and 1.774 million and a latest four-week average of 1.779 million. The stable 206,000 initial-claims print reduces near-term upside pressure. I found no federal extended-benefit or policy mechanism in the DOL release that should mechanically shift regular state-program seasonally adjusted insured unemployment.

Prior/update/interval: persistence prior = latest official CCSA first print of 1.774 million for week ending 2026-08-29; historical sample = fetched DOL same-variant weekly CCSA first-print levels from 2026-03-14 through 2026-08-29, using successive weekly changes. Adjustment components = +0.003 million toward the latest four-week average of 1.779, +0.002 million for possible post-Labor-Day reaccumulation, and +0.000 million policy-mechanism effect, giving point = 1.774 + 0.003 + 0.002 = 1.779 million. Successive-change dispersion over the fetched sample gives one-week sigma = 0.0195 million; for the three-week horizon to 2026-09-19, assuming independent weekly changes, sigma = sqrt(3)*0.0195 = 0.0337 million; 80% half-width = 1.28*sigma = 1.28*0.0337 = 0.043 million. Final implied bounds are 1.779 +/- 0.043 = [1.736, 1.822].

Counter-considerations: upside risk is a sequence of weak reemployment, delayed post-holiday filings, or longer benefit duration that would push CCSA above 1.822 million. Downside risk is sustained low initial claims and faster exits from insured unemployment, which would land below 1.736 million. Outside the interval would likely require a clear break from the recent 1.77 to 1.82 million range, not just ordinary one-week noise.

Update from prior run: the previous public Thesis continued-claims run for week ending 2026-09-12 used 1.779 million through 2026-08-22 and forecast 1.785 million. The new DOL 2026-09-10 print added week ending 2026-08-29 at 1.774 million and revised 2026-08-22 to 1.775 million, so I move the point down to 1.779 while keeping a similar volatility-based interval.

Review disposition: accepted the resolver critique by making the target-registration timing discrepancy explicit while preserving the registered slug, unit, dataPointId, source URL, and resolutionDate. Accepted the optional clarifications by naming the 2026-08-29 first-print sample endpoint and stating the independent weekly-change assumption behind the sqrt(3) horizon scaling.

calibrated forecast · 80% CI
1.8M[1.7M · 1.8M]
Target metadata

Data point: dol.eta.continued_claims.sa.week_2026-09-19.first_print

More government data forecasts