US initial claims, week ending Jul. 4 2026

The target is the first-print DOL/ETA advance seasonally adjusted initial claims count for the week ending July 4, 2026, reported in thousands. The canonical catalog slug and unit match the target contract.

Base-rate/reference-class anchor: the tightest reference class is recent weekly DOL first-release-adjacent prints for the same SA claims series. The latest six official table values average 220.8 thousand and the latest four-week average is 222.0 thousand, while the latest value is 215.0 thousand.

Level, momentum, and mechanism: the level is low but not unusually low relative to spring 2026. Momentum is downward from the June 6 peak of 230 to 215, while continuing claims and the weak payroll print suggest slower hiring rather than a clean layoff surge. The July 4 holiday week adds seasonal-factor risk, so persistence alone is too low.

Prior/update/interval: persistence prior is the latest DOL print of 215 thousand blended with the recent six-week mean of 220.8 and four-week average of 222.0. I add 3 thousand for mean reversion toward the recent average, 1 thousand for weak payroll and continuing-claims pressure, and 1 thousand for holiday-week seasonal-adjustment upside risk, giving 220 thousand. The 80% interval uses recent realized weekly dispersion from the official table, where adjacent changes included +13, +5, -3, -11, and -1 thousand, widened for holiday-week noise to roughly -16/+18 thousand.

Recent six-week mean = (212 + 225 + 230 + 227 + 216 + 215) / 6 = 220.8 thousand. Point = 215 latest + 3 mean-reversion + 1 labor-softening + 1 holiday-seasonal risk = 220 thousand. Interval = 220 - 16 = 204 and 220 + 18 = 238 thousand.

Counter-consideration and scenarios: downside outside the interval would be a clean continuation of the last two weekly declines plus favorable seasonal adjustment, pushing below 204 thousand. Upside outside the interval would be holiday-week filing disruption, auto-sector layoffs, or a broader layoff pickup pushing above 238 thousand. The central case is a rebound toward the recent four-week average without a break in layoff behavior.

Framing and exact resolver: this is the Department of Labor ETA Unemployment Insurance Weekly Claims advance seasonally adjusted initial claims series, in thousands, for the week ending July 4, 2026. The target is the first print only; later revisions are ignored. The variant is seasonally adjusted initial claims, not NSA claims, continuing claims, or four-week average.

Base rate/reference class: the 2026 weekly SA initial-claims reference class from January 3 through June 27 runs mostly in a 190 to 230 thousand range, with the latest four-week average at 222 thousand. Persistence from 215 is the main base rate, but the gap to the four-week average and holiday-week noise argue against forecasting another sharp decline.

Level, momentum, and mechanism split: level is low relative to the June four-week average; momentum is downward over June 6 to June 27; continuing claims at 1.814 million point to slower hiring rather than mass layoffs; the July 4 week can distort seasonal factors, so I add only a small rebound from the latest print.

Prior/update/interval: persistence prior from the latest official SA print is 215 thousand, historical sample is weekly SA initial-claims values from Jan. 3 to Jun. 27 2026, adjustment components are +2 thousand mean reversion toward the 222 thousand four-week average and +1 thousand holiday/seasonal-noise allowance, giving point 218. For interval, compute successive weekly changes from the fetched 2026 history; sigma = 10.66 thousand, so 1.28*sigma = 13.64 thousand. Rounded to whole thousands, 218 - 14 = 204 and 218 + 14 = 232.

Counter-consideration and falsification: upside risk is a holiday-week seasonal-adjustment miss or state-level layoff bulge that would land above the interval, over 232 thousand. Downside risk is continued low layoff activity plus favorable seasonal adjustment that would land below the interval, under 204 thousand. The central case is a modest rebound, not a break in the labor market.

The resolver is the DOL/ETA advance seasonally adjusted initial-claims count for the week ending July 4, 2026. All claims anchors use the same DOL seasonally adjusted initial-claims variant, not NSA claims, continued claims, FRED, or later revised values.

Base rate/reference class: the tightest reference class is recent DOL first-release-adjacent seasonally adjusted initial claims. The latest 10 same-variant values average 213.6 thousand, while the latest four-week average is 222 thousand and the latest weekly value is 215 thousand. That puts the base rate near the high teens rather than at the early-June peak.

Level, momentum, one-off, and policy split: the current level is low relative to late May and early June, momentum is downward over the last three prints from 230 to 227 to 216 to 215, and the four-week average is also easing. The main one-off is the Independence Day week seasonal-adjustment problem; the broader labor market from BLS is softer, but there is no public policy mechanism in the checked material that should create a discrete claims break for this week.

Prior/update/interval: persistence prior is latest-value persistence at 215 thousand, with a reference class of the 10 fetched same-variant DOL values from April 25 through June 27. I add +2 thousand for mean reversion toward the 222 thousand four-week average, +1 thousand for weak payroll and downward revision risk, and +1 thousand for holiday-week seasonal volatility, giving 219 thousand. Interval method uses realized dispersion of the fetched same-variant weekly flow values themselves; sigma = 12.4 thousand, so 1.28*sigma = 15.9 thousand. I round the half-width to 16 thousand because this is within the empirical 80% scale and the holiday week argues against narrowing.

Point calculation: 215 latest + 2 mean-reversion adjustment + 1 weak-labor-market adjustment + 1 holiday-volatility adjustment = 219 thousand. Interval calculation: sigma = 12.4 thousand; 1.28*sigma = 15.9 thousand; rounded half-width = 16 thousand, so 219 - 16 = 203 and 219 + 16 = 235 thousand.

Counter-consideration: upside risk is that holiday-week seasonal adjustment and softer payroll demand lift the first print above 235 thousand. Downside risk is that low layoffs persist after the 215 thousand latest print and claims land below 203 thousand. An outside the interval result would most likely come from holiday timing, state processing backlogs, or a sharp new layoff shock not visible in the latest official claims table.

Framing and exact resolver: this is the DOL ETA advance seasonally adjusted initial claims series for the week ending July 4, 2026, reported in thousands on the first UI Weekly Claims release. The variant is seasonally adjusted initial claims, not NSA claims, continuing claims, or the 4-week average.

Reference class base rate: recent DOL seasonally adjusted weekly claims sit in the low-200-thousand range, with the latest four revised or advance levels 230, 227, 216, and 215 thousand. That favors persistence around 215 to 222 rather than a break above 230 unless the holiday week produces a seasonal-adjustment miss.

Level, momentum, and mechanism: level is low but not unusually low relative to spring 2026; momentum is mildly downward after the June peak, while the 4-week average at 222 is above the latest print. The week includes Independence Day, so seasonal factors and employer shutdown timing add noise, but the target is SA and should remove most regular holiday seasonality.

Prior/update/interval: persistence prior uses the latest official SA first/revised level of 215 thousand and a recent-reference-class center between latest 215 and 4-week average 222; adjustment components are +1 thousand for mean reversion toward the 4-week average, +0 for labor-market momentum after offsetting the three-week decline, and +0 for holiday effects after seasonal adjustment, giving point 216. For interval width, I used the DOL table's 53 successive weekly SA changes from June 2025 through June 2026; sigma = 10.42 thousand, so 1.28*sigma = 13.34 thousand, rounded to a 13-thousand half-width. Thus 216 - 13 = 203 and 216 + 13 = 229.

Counter-consideration: upside risk is a holiday-week seasonal adjustment miss or a fresh layoff wave lifting claims above 229 thousand; downside risk is continued normalization after the June bump or fewer auto-related layoffs, which would land below 203 thousand. Outside the interval would most likely require a one-week seasonal-factor surprise rather than a gradual labor-market signal.

Framing and exact resolver: the target is the DOL Unemployment Insurance Weekly Claims advance seasonally adjusted Initial Claims figure for regular state programs, in thousands, for the week ending July 4, 2026, first print only. The variant is seasonally adjusted, not NSA, and all anchors below use the same SA initial-claims variant unless explicitly described as release-schedule context.

Reference class and base rate: for a one-week-ahead SA initial-claims print, the most relevant reference class is the DOL weekly SA history in the current release table. Over the fetched 54 weekly values from June 21, 2025 through June 27, 2026, levels mostly sit around 200k-236k, with the latest four weeks 230k, 227k, 216k, and 215k and the latest 4-week average at 222k. The base rate is therefore a near-flat next print around the latest level to recent average, before holiday-week and noise adjustments.

Level, momentum, one-off, and policy mechanisms: level anchors at 215k latest and 222k four-week average; momentum is slightly down over the last two weeks, from 227k to 216k to 215k; one-off risk is elevated because the target week includes the July 4 holiday and seasonal factors can be imperfect; policy mechanisms do not point to an abrupt claims regime change by this release.

Prior/update/interval: persistence prior model starts from latest SA initial claims of 215k with the 4-week-average anchor at 222k and comparable 2025 early-July value of 228k; historical sample is the fetched DOL SA weekly table from June 21, 2025 through June 27, 2026. Adjustment components: +2k toward the recent average, -1k for short-term downward momentum, +1k for early-July seasonal-noise skew, giving a center near 217k. From successive weekly changes in the fetched table, sample sigma = 10.4 thousand, so 1.28*sigma = 13.3 thousand. A symmetric 80% interval around 217k would be about 204k to 230k; the ladder-implied 80% bounds are 204k to 231k, essentially the same width and not widened beyond the realized-dispersion reference.

Ladder: P(X <= 198) = 0.04; P(X <= 203) = 0.09; P(X <= 208) = 0.18; P(X <= 212) = 0.30; P(X <= 215) = 0.42; P(X <= 218) = 0.53; P(X <= 221) = 0.64; P(X <= 224) = 0.74; P(X <= 228) = 0.84; P(X <= 232) = 0.91; P(X <= 236) = 0.96; P(X <= 242) = 0.98. Linear interpolation gives the 10th percentile between 203 and 208 at 203.6, rounded to 204; the median between 215 and 218 at 217.2, rounded to 217; and the 90th percentile between 228 and 232 at 231.4, rounded to 231.

Counter-considerations: upside risk is a rebound in filings after holiday timing or state processing backlogs, especially if the latest 215k print understated the run rate; that would land above the interval if the print exceeds 231k. Downside risk is another unusually clean SA print with broad state declines or front-loaded holiday processing before the reference week, which would land below the interval if the print is under 204k. Outside the interval is most likely from seasonal-adjustment error around the July 4 week rather than from a true labor-market break.

Review disposition: accepted the provenance critique by making the stable DOL claims archive the resolution URL while retaining the current-release PDF and claims data page in sourceContext; accepted the clarification that the 2025 comparison is the July 5, 2025 early-July week; rejected no substantive forecast recalibration because the critique did not change the fetched SA evidence or ladder arithmetic.

Derived run: the pointwise median of the CDFs of 3 independent thesis.analyst fast rollouts on this target — median prediction sampling per Turtel et al. 2025 (arXiv:2505.17989). No new model call; this run is a deterministic aggregate of the recorded rollouts at 2026-07-08T02:44:18Z, 2026-07-08T02:44:20Z, 2026-07-08T02:44:44Z. Drivers and resolver fields mirror the rollout closest to the median.

Median CDF quantiles: q10 = 203.0, q50 = 218.0, q90 = 232.0. Constituent points [218, 219, 216] with 80% widths [28, 32, 26]; the median interval inherits the central rollout mass rather than averaging tails.