§US initial claims forecast for week ending 2026-08-29
Framing and exact resolver: the target is U.S. Department of Labor Employment and Training Administration initial unemployment insurance claims, seasonally adjusted, advance first print, series ICSA, for week ending 2026-08-29, in thousands. The registered target uses ALFRED ICSA as the mechanical advance-vintage adapter while DOL remains the official agency source. The DOL archive says the weekly claims news release is published each Thursday at 8:30 a.m. Eastern except listed holiday exceptions; for this target week that points to the 2026-09-03 DOL release, while the registered sourceBinding window is 2026-09-07 to 2026-09-11, so I keep the top-level resolutionDate tied to the registered bound and state the calendar discrepancy rather than changing the target.
Fetched the current DOL UI Weekly Claims PDF at https://www.dol.gov/ui/data.pdf.↳ The 2026-08-20 DOL release reported that for week ending 2026-08-15 the advance seasonally adjusted initial claims figure was 206,000, down 6,000 from the previous week's revised 212,000; the same release table listed 2026-08-08 at 212,000 revised and 2026-08-01 at 200,000 revised.Fetched archived DOL releases for 2026-08-13, 2026-08-06, and 2026-07-30.↳ The 2026-08-13 release first printed week ending 2026-08-08 at 209,000; the 2026-08-06 release first printed week ending 2026-08-01 at 199,000; the 2026-07-30 release first printed week ending 2026-07-25 at 197,000.Fetched archived DOL releases for 2026-07-23, 2026-07-16, 2026-07-09, and 2026-07-02.↳ The 2026-07-23 release first printed week ending 2026-07-18 at 187,000; the 2026-07-16 release first printed week ending 2026-07-11 at 208,000; the 2026-07-09 release first printed week ending 2026-07-04 at 215,000; the 2026-07-02 release first printed week ending 2026-06-27 at 215,000.Base rate / reference class: the recent official-source reference class is eight advance seasonally adjusted initial-claims first prints from 2026-06-27 through 2026-08-15: 215, 215, 208, 187, 197, 199, 209, and 206 thousand. The base rate center is about 205 thousand by mean and 207 thousand by median, with the latest first print at 206 thousand.
Level, momentum, one-off, and policy effects: level anchors near 206 thousand; momentum is mixed after 187 to 197 to 199 to 209 to 206; mean reversion points only slightly toward the 205 to 207 center; no public policy mechanism in the claims release suggests a discrete jump. The series variant is advance seasonally adjusted initial claims, not NSA claims, continued claims, the insured unemployment rate, or the four-week moving average.
Prior/update/interval: prior = recent first-print persistence/reference-class model using the fetched official DOL ICSA prints from 2026-06-27 through 2026-08-15; adjustment components = 206 latest level, -1 thousand mild pull toward the eight-print mean of 204.5, and 0 thousand for policy mechanism, giving point = 205. Interval method = realized dispersion of successive changes in the same first-print history, transparent but sample-limited to seven week-to-week changes. Changes were 0, -7, -21, +10, +2, +10, -3 thousand; sample sigma = 10.73 thousand; one-week 80% half-width = 1.28*sigma = 13.73 thousand. Because the target week is two unreleased weekly transitions after the latest print, widen by sqrt(2): 13.73*1.414 = 19.42 thousand. Final implied bounds = 205 +/- 19.42 = [185.58, 224.42], rounded to [186, 224].
Counter-considerations: upside risk is a renewed layoff pulse or another adverse seasonal-adjustment surprise after the 187 thousand July trough, which would land above the interval if the first print exceeds 224 thousand. Downside risk is continued very low late-summer layoffs with favorable seasonal factors, which would land below the interval if the first print is under 186 thousand. Outside the interval would require a move larger than the recent two-week noise without a currently observed mechanism.
Review disposition: accepted the reviewer suggestion to state that the volatility estimate is based on only seven week-to-week changes, and accepted the resolver-clarity suggestion by removing an unnecessary rounding clause from the resolution rule. No required fixes were identified.