US initial unemployment claims, week ending 2026-06-20 (thousands)

Same series as the prior cell, one week further out. With no information yet on the 6/13 print, the best estimate reverts slightly further toward the trailing mean, and the interval is a touch wider to reflect the longer horizon.

Point: pull a bit closer to the trailing-8 mean (211.5k) than the 6/13 cell, landing at 220k (vs 222k for the nearer week). 80% CI from WoW vol over a ~2-week-ahead path: half-width = 1.28*10.9*sqrt(1.2) ~ 15.3k, widened to 22k. Interval = 220 +/- 22 = [198k, 242k].

What lands it outside [198, 242]: a fresh layoff cluster or seasonal-adjustment overshoot above 242k, or an unusually clean low print (e.g. holiday-week filing dip) below 198k. The further horizon means a 6/13 surprise would also shift the 6/20 baseline.

The resolver is the DOL ETA first-print advance seasonally adjusted initial claims count for the week ending Saturday, June 20, 2026, reported in the Unemployment Insurance Weekly Claims Report. The DOL economic-data page identifies the ETA Office of Unemployment Insurance as the publisher of this weekly report and links the current official PDF.

Base-rate/reference-class: weekly initial claims are noisy but mean-reverting at a 1-3 week horizon. Recent 2026 levels cluster near 210,000-229,000, while the same June window in 2025 ran 231,000-246,000, so a central forecast should stay near the latest value but not fully extrapolate the late-May jump.

Recent 5-week average using DOL/FRED values is (212,000 + 210,000 + 212,000 + 225,000 + 229,000) / 5 = 217,600. The latest 4-week average is 219,000. A same-season 2025 anchor for week ending June 21 is 236,000, but current June 6 is 17,000 below the 2025 comparable 246,000, implying an adjusted seasonal anchor near 219,000. I blend latest 229,000, recent-average 218,000, and seasonal-adjusted 219,000, then allow for another school-calendar noisy week to set the point at 226,000.

Counter-consideration: the June 6 unadjusted actual claims jumped by 39,713 and state moves in California, Minnesota, Pennsylvania, Texas, and other large states show timing noise that could persist into the June 20 week; this raises upside risk even though the seasonally adjusted series may mean-revert.

For an 80% interval, recent weekly moves include +13,000, +4,000, -25,000, and +15,000 over spring 2026, and June school-season volatility can add several thousand more. Around a 226,000 point, a practical 80% range is roughly -21,000/+24,000, giving 205,000 to 250,000.

The control run partially fades the latest claims print toward the trailing mean with symmetric weekly noise.

The pack run keeps more of the recent increase because labor-market cross-checks do not yet show enough softening to force a snap-back.

The resolver is the first-print national seasonally adjusted initial claims figure in the ETA weekly claims report for the week ending June 20, 2026; the unit here is thousands of claims.

Base-rate reference class: one-week changes in this series are usually small relative to the level, and the latest sequence 210k, 212k, 225k, 230k, 226k puts a neutral forecast in the low-to-mid 220s rather than implying a new break.

Recent anchors: 3-week average = (226 + 230 + 225) / 3 = 227.0; 5-week average = (226 + 230 + 225 + 212 + 210) / 5 = 220.6; latest 4-week average = 223.25. I put the point at 225, slightly below the 3-week average but above the 5-week average because early-June claims have been elevated versus May.

Counter-consideration: the end-of-school-year period can create seasonal-adjustment noise, so a sharper move toward 235k or back toward 210k is plausible even without a real layoff-cycle shift.

For an 80% interval, I use roughly +/-17k around the 225k point, reflecting normal weekly volatility and recent early-summer noise: low 208, high 243.