JOLTS job openings, May 2026 (first print)

Computed MoM change in openings: last 12 changes (thousands) = [212,-106,-115,-170,250,1,-324,-296,690,-318,-35,731]. Population stdev of the monthly change over the last 24 months = 320k. This is the single most volatile of the six series; an 80% band (z=1.28) implies ~+/-410k from the conditional centre, before mean-reversion.

No persistent trend; the 6-month mean change is only +75k and the 12-month mean +43k. The series is range-bound 6.5-7.6M. The dominant feature for May is the April +731k spike, which is an outlier likely to partly reverse.

Base rate over the last 12 months: mean level ~7.04M, with openings printing in the high-6M / low-7M range far more often than near 7.6M. Unconditionally the centre of gravity is ~7.0-7.2M, below April's 7.618M.

After the +690k jump in Sep->... and the +731k April jump, history shows large positive shocks are typically followed by give-back (e.g. the +690k in Jan 2026 was followed by -318k). Applying partial reversion (~40-50% of the +731k) to April's 7.618M lands May near ~7.25-7.40M.

Outside [6.85, 7.85] if: April's 7.618M is revised up and May extends the surge above ~7.85M; or a demand-side pullback (oil shock, hiring freeze) snaps openings back below ~6.85M as in late-2025 (Dec 6.55M). JOLTS' low response rate makes a >400k surprise routine.

Partial mean-reversion from the April spike toward the 7.0-7.2M base, shaded up for macro strength -> point 7.35M. 80% CI = 7.35 +/- ~0.50M -> [6.85, 7.85].

The control treats the April openings surge as a noisy high print and pulls the center back toward the 6.9M to 7.2M recent range.

Payroll and claims cross-checks make a full reversal of April's openings spike less attractive than the no-pack control, but the JOLTS pack keeps a wide interval for response-rate and revision noise.

The target is the first official BLS JOLTS print for seasonally adjusted total nonfarm job openings in May 2026, reported in thousands; later revisions do not count.

Base-rate/reference class: month-to-month JOLTS job openings are noisy and often revise, so I anchor on the recent five-month range of 6,550 to 7,618 thousand and the latest three-month average near 7,142 thousand rather than extrapolating the full April jump.

The upside case is that solid May payroll growth and stable unemployment mean labor demand remained firm, so April's 7,618 thousand could persist into May.

Counter-consideration: April's 731 thousand openings increase was unusually large while hires fell to 5,116 thousand, so some mean reversion or measurement noise is plausible in the first May print.

Anchor around Apr 2026 7,618 and recent three-month average (7,618 + 6,887 + 6,922) / 3 = 7,142; weighting April persistence against mean reversion gives about 7,450 thousand. An 80 percent interval of 6,850 to 8,050 allows roughly +/-600 thousand, covering typical JOLTS volatility and the April spike risk.