US real average hourly earnings MoM, July 2026

Framing and exact resolver: this is the BLS Real Earnings Table A-1 series for real average hourly earnings for all employees on private nonfarm payrolls, seasonally adjusted, over-the-month percent change. The table says CPI-U is used to deflate the earnings series, so all anchors here use the same SA all-employees private payroll earnings variant and CPI-U deflator.

Reference class and base rate: the directly fetched real-AHE over-the-month reference class from the current Table A-1 is 0.4, -0.6, -0.5, and -0.1 percent, averaging -0.2 percent; the recent base rate is therefore negative real wage growth when headline CPI is running above the nominal hourly earnings pace.

Current-release adjustment: nominal wage momentum is still near 0.3 percent monthly, using June Table B-3's $37.64 versus $37.51 as a live wage anchor. CPI momentum is less favorable, with the latest all-items CPI-U monthly gains 0.9, 0.6, and 0.5 percent, but some May energy pressure could partly mean-revert by July. Combining a July nominal AHE assumption near +0.30 percent with a CPI-U assumption near +0.35 to +0.40 percent points to a small negative real hourly earnings print.

Prior/update/interval: persistence prior is the recent BLS Table A-1 real-AHE MoM base rate using fetched values [0.4, -0.6, -0.5, -0.1], mean = -0.20. Adjustment components are +0.10 for steady nominal wage growth near 0.3% and -0.05 for CPI still elevated but not as extreme as May, giving point near -0.15, rounded to -0.1. Interval method uses realized dispersion of the fetched change values themselves because this is a change series and the small four-observation sample makes the 80% interval deliberately approximate: sigma = 0.45, so 80% half-width is roughly 1.28*sigma = 1.28*0.45 = 0.58; around -0.1 this gives about [-0.68, 0.48], rounded to [-0.7, 0.5].

Counter-consideration: upside risk is a July CPI relief print, especially an energy reversal, combined with another 0.3-0.4 percent nominal wage month, which would land above the interval. Downside risk is another gasoline or broad services CPI spike with only 0.2 percent nominal wage growth, which would land below the interval. Outside the interval would require roughly real AHE above +0.5 percent or below -0.7 percent on the first print.

Review disposition: accepted the optional clarification that the interval volatility estimate is based on only four recent Table A-1 observations and is approximate; no required fixes were raised, and the BLS schedule source was already included in sourceContext and the calendar tool step.

The target is the seasonally adjusted all-private-employees series in BLS Real Earnings Table A-1, resolved to the first July 2026 print on the official release page without later revision.

The resolution variant is the gross real average hourly earnings measure for all employees on private nonfarm payrolls, seasonally adjusted, rather than the production-and-nonsupervisory or unadjusted variant. The relevant table is BLS Table A-1.

The base rate/reference class is the five available 2026 monthly first-print changes before the target: mostly small positive or negative moves, with a recent mean of -0.16 percent. June nominal wage growth of 0.3 percent suggests a modest positive real result if July CPI inflation is contained, while the recent inflation-adjusted pattern argues against a large increase.

A persistence-and-current-momentum model combines the recent real-change mean with a modest upward adjustment for the latest 0.3-percent nominal wage increase, producing a central estimate near 0.1 percent.

Prior/update/interval: the persistence prior uses the five fetched January-May 2026 Table A-1 changes [0.3, 0.1, -0.6, -0.5, -0.1]; the current update is nominal wage momentum of 0.3 percent and an assumed near-offset from July CPI; sigma = 0.34 percentage points, calculated as the population standard deviation of those five change values; the 80% half-width is roughly 1.28*0.34 = 0.44, rounded to one decimal, giving final implied bounds of -0.3 to 0.5 percent.

Upside risk is a July CPI increase below the recent wage pace, which would land above the interval if real hourly earnings rise at least 0.6 percent. Downside risk is a CPI reacceleration or weaker nominal wage growth; a real decline of -0.4 percent or less would land outside the interval.

Point estimate = 0.1 percent. Using sigma = 0.34 and 1.28*sigma = 0.44, the one-decimal 80% interval is 0.1 - 0.4 = -0.3 percent to 0.1 + 0.4 = 0.5 percent.

The target is the first-print, seasonally adjusted Table A-1 value for all employees on private nonfarm payrolls, resolved from the BLS Real Earnings release without later revisions. The BLS release calendar verifies July 2026 Real Earnings publication on August 12, 2026.

The base rate is centered near zero but has recently shifted negative: four of the five available 2026 monthly observations were at or below zero, while nominal earnings growth provides a countervailing positive force. The relevant variant is gross real average hourly earnings for all private nonfarm employees, seasonally adjusted, specifically BLS Table A-1.

The persistence prior points to a small negative July result because April and May were -0.5% and -0.1%, while the June nominal wage increase limits the downside absent another inflation surge. I therefore use -0.1% as the modal one-decimal outcome.

Prior/update/interval: The persistence prior is the five-point January-May 2026 reference class [0.3, 0.2, -0.6, -0.5, -0.1]. For this change series, sigma is computed from the values themselves: mean = -0.14, sigma = 0.36, and the nominal 80% half-width is roughly 1.28*sigma = 0.46. Rounding to the BLS one-decimal unit gives an interval of -0.6% to 0.4%; the recent inflation shock and release-to-release volatility justify retaining that full width.

Downside risk is a July CPI acceleration combined with weak nominal wage growth, which could produce a result below -0.6%. Upside risk is softer inflation with another 0.3% or larger nominal wage increase, which could land above 0.4%; those are the concrete cases outside the interval.

Point estimate = -0.1%. Using sigma = 0.36 and 1.28*sigma = 0.46, the rounded 80% interval is approximately -0.1% - 0.5% to -0.1% + 0.5%, or [-0.6%, 0.4%].

The target is the first-print, one-decimal, seasonally adjusted Table A-1 value for all employees on private nonfarm payrolls. The BLS schedule places July Real Earnings on August 12, 2026 at 08:30 ET, verifying the resolution date.

The base rate/reference class is the four available 2026 Table A-1 monthly changes, whose mean is -0.25%. The June nominal wage increase supports some improvement, but real earnings still depend on the not-yet-released July CPI and recent real outcomes have been weak.

Level, momentum, and one-off effects are separated as follows: the level is modestly positive nominal wage growth; momentum is the recent run of -0.6%, -0.5%, and -0.1% real changes; the main one-off risk is monthly CPI volatility; the policy mechanism is ordinary BLS deflation of nominal earnings by CPI-U, with no policy-rate conditioning.

For the change series, fetched values themselves are used: [0.2, -0.6, -0.5, -0.1]. Their population standard deviation is sigma = 0.32 percentage points, so the 80% half-width is roughly 1.28*0.32 = 0.41. Centering near -0.2 gives an interval approximately [-0.6, 0.2], close to the dispersion-based width.

Upside risk is a July nominal wage gain near 0.5% or softer CPI, which would land above the interval. Downside risk is nominal wage growth near 0.1% combined with CPI inflation near 0.6% or more, which would land below the interval.

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 2026-07-10T15:12:31Z, 2026-07-10T15:15:40Z, 2026-07-10T15:18:50Z. Drivers and resolver fields mirror the rollout closest to the median.

Median CDF quantiles: q10 = -0.6, q50 = -0.1, q90 = 0.4. Constituent points [0.1, -0.1, -0.2] with 80% widths [0.8, 1.0, 0.8]; the median interval inherits the central rollout mass rather than averaging tails.

The resolver is BLS Real Earnings Table A-1: the seasonally adjusted over-the-month percent change in real average hourly earnings for all employees on private nonfarm payrolls. This is the all-employees, CPI-U-deflated variant, not the production-and-nonsupervisory CPI-W series; resolve only to the July first print, rounded to BLS's one decimal, with no later revision substitution.

The five-observation reference class has a base rate of -0.16 percent per month: January through May were 0.3, 0.1, -0.6, -0.5, and -0.1 percent. The latest two releases show real pay falling as CPI-U outran modest nominal wage gains, so the baseline remains slightly negative.

Prior/update/interval: Persistence prior is the January-May 2026 first-print Table A-1 change sample [0.3, 0.1, -0.6, -0.5, -0.1], with mean = -0.16 and sample sigma = 0.38 percentage points. The adjustment components are a +0.1 point pull from June's 0.3 percent nominal-pay growth, offset by continued CPI-U deflator uncertainty and no assumed July CPI slowdown, yielding -0.1 percent. The 80% half-width is 1.28*0.38 = 0.49 percentage points; rounded to the release's one-decimal precision, implied bounds are -0.6 to 0.4 percent.

upside risk: a July CPI-U print near zero alongside another roughly 0.3 percent nominal hourly-pay gain would lift real earnings above 0.4 percent. downside risk: a CPI-U increase materially above nominal wage growth, or a weak/negative nominal-pay print, would land below -0.6 percent. A sharp energy-driven CPI move or unusual seasonal-factor outcome is the main outside the interval scenario.

The resolver is BLS Real Earnings Table A-1: the seasonally adjusted over-the-month percent change in real average hourly earnings for all employees on private nonfarm payrolls. Anchors and history below use this same Table A-1 all-employees, seasonally adjusted variant; this target resolves to its first printed one-decimal value, without later revision.

The reference class is monthly, seasonally adjusted Table A-1 real hourly-earnings changes. The recent base rate is slightly negative after January through May values of 0.3, 0.2, -0.6, -0.5, and -0.1 percent, but the most recent two observations show the contraction decelerating.

Prior/update/interval: I use a zero-centered persistence prior with the five fetched Table A-1 change observations (0.3, 0.2, -0.6, -0.5, -0.1). Their mean is -0.14 and population sigma = 0.36 percentage points; 1.28*sigma = 0.46 percentage points. I update the prior to 0.1 percent because May's -0.1 followed the earlier -0.6 and -0.5 declines, while the nominal-wage-minus-CPI mechanism can return a modest positive reading. Rounding the approximately 0.46-point half-width to the publication grid gives an 80% interval of -0.4 to 0.6 percent.

Upside risk is a July nominal-pay acceleration combined with subdued CPI-U, which would land above the interval if real hourly earnings rise more than 0.6 percent. Downside risk is another inflation overshoot or unfavorable high-wage/low-wage payroll-composition shift; a sufficiently large combination would land below the interval.

The resolver is BLS Real Earnings Table A-1, not the production-and-nonsupervisory Table A-2: the required variant is all employees on private nonfarm payrolls, seasonally adjusted, over-the-month percent change in real average hourly earnings. Table A-1 uses CPI-U as its deflator.

Base rate/reference class: the available eleven monthly Table A-1 changes average about -0.07 percent and are centered near zero, with recent March-through-May inflation outpacing nominal hourly-pay growth. The June 0.3 percent nominal wage increase offsets part of that recent weakness but does not by itself imply a positive real print.

Prior/update/interval: persistence prior is the available May 2025-May 2026 Table A-1 monthly-change sample [0.4,-0.1,0.1,0.1,-0.1,-0.3,0.3,0.0,-0.6,-0.5,-0.1], with mean -0.07 percent; adjustment components are June's 0.3 percent nominal hourly-pay momentum, likely near-matching CPI-U inflation, and composition/seasonal noise. For this change series, sigma = 0.31 percentage points (sample standard deviation of the fetched values); 1.28*sigma = 1.28*0.31 = 0.40 percentage points. Rounding the implied -0.07 percent center to the release's one-decimal convention gives -0.1 percent and bounds of -0.5 to 0.3 percent.

Upside risk: a July nominal-pay gain around 0.4 percent with CPI-U inflation near 0.1 percent would land at about 0.3 percent. Downside risk: a renewed CPI-U increase near 0.7 percent alongside only 0.1 percent nominal wage growth would land below the interval, around -0.6 percent. A large industry-composition shift could also put the first print outside the interval.

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 2026-07-10T15:39:47Z, 2026-07-10T15:43:51Z, 2026-07-10T15:47:55Z. Drivers and resolver fields mirror the rollout closest to the median.

Median CDF quantiles: q10 = -0.5, q50 = -0.1, q90 = 0.4. Constituent points [-0.1, 0.1, -0.1] with 80% widths [1.0, 1.0, 0.8]; the median interval inherits the central rollout mass rather than averaging tails.

Framing and exact resolver: this forecast is for BLS Real Earnings Table A-1, all employees on private nonfarm payrolls, seasonally adjusted over-the-month percent change in real average hourly earnings, July 2026 first print, in percent and rounded to the one-decimal value printed by BLS. The target uses the same SA real earnings variant throughout.

Base rate/reference class: for this change-rate series, the direct recent official reference class is the contiguous BLS Table A-1 over-the-month real average hourly earnings sample for Mar-May 2026: -0.6, -0.5, and -0.1 percent. That anchors the distribution near negative but improving, before adjusting for current nominal wage momentum and July inflation risk.

Prior/update/interval: persistence prior is the contiguous recent Table A-1 change-rate sample [-0.6, -0.5, -0.1], with mean -0.4. For a change/flow target I compute sigma from the values themselves: squared deviations from -0.4 are 0.04, 0.01, and 0.09; sample variance = 0.14/2 = 0.07, so sigma = 0.26 and the normal 80% half-width is roughly 1.28*sigma = 0.33. Current-release adjustment components are +0.2 from steady nominal AHE around 0.3 percent, +0.1 from expected partial energy/CPI moderation versus the March-May spike, and +0.1 because the three-month sample is unusually inflation-heavy and too small for full-cycle volatility, moving the center from -0.4 to about 0.0. The ladder-implied 80% half-width is 0.5, about 1.52x the 1.28*sigma width, widened for the small sample, CPI energy volatility, and one-decimal first-print rounding. No longer time-series model is used because the target is dominated by near-term CPI and nominal wage inputs.

Ladder: P(X <= -0.8) = 0.04; P(X <= -0.6) = 0.08; P(X <= -0.5) = 0.11; P(X <= -0.4) = 0.16; P(X <= -0.3) = 0.23; P(X <= -0.2) = 0.32; P(X <= -0.1) = 0.42; P(X <= 0.0) = 0.54; P(X <= 0.1) = 0.65; P(X <= 0.2) = 0.75; P(X <= 0.3) = 0.82; P(X <= 0.4) = 0.87; P(X <= 0.6) = 0.93; P(X <= 0.8) = 0.97; P(X <= 1.0) = 0.99. Linear interpolation gives p10 at -0.533, p50 at -0.033, and p90 at 0.500; rounded to BLS one-decimal print precision these are -0.5, 0.0, and 0.5.

Counter-considerations: upside risk is a July CPI slowdown toward 0.0 to 0.1 percent with nominal AHE still near 0.3 percent, which could put real AHE near or above 0.4 percent. Downside risk is another gasoline or energy-services CPI jump alongside softer hourly earnings, which could push the print below -0.4 percent. A large energy reversal plus a strong wage mix effect would land above the interval; another CPI shock above roughly 0.8 percent with weak nominal earnings would land below the interval.

Review disposition: accepted the critique that the base-rate sample should not mix May 2025 with recent 2026 prints, so the prior and sigma now use the contiguous Mar-May 2026 Table A-1 values and label their small-sample limitation; accepted the optional note to state that no longer time-series model was used.

Framing and exact resolver: this forecast is for the seasonally adjusted over-the-month percent change in real average hourly earnings for all employees on private nonfarm payrolls in BLS Real Earnings Table A-1, first print for July 2026, using the table's one-decimal percent value and no later revisions. The variant is the same throughout: all employees, private nonfarm payrolls, seasonally adjusted, deflated by CPI-U in Table A-1.

Base rate/reference class: the immediate same-table reference class is the latest three published real hourly m/m changes, which average -0.4 percent. That base rate is depressed by unusually firm CPI-U prints of 0.9, 0.6, and 0.5 percent, while nominal hourly earnings remained roughly 0.2 to 0.3 percent per month.

Prior/update/interval: persistence prior uses the latest same-variant Table A-1 real average hourly earnings m/m values [-0.6, -0.5, -0.1], mean = (-0.6 - 0.5 - 0.1) / 3 = -0.4. I adjust +0.3 percentage point for July because nominal hourly earnings were still near +0.3 percent in June and CPI is unlikely to repeat the 0.5 to 0.9 percent monthly pace from Mar-May, giving point = -0.4 + 0.3 = -0.1. For realized dispersion on this change series, sample sigma = sqrt((( -0.6 + 0.4)^2 + (-0.5 + 0.4)^2 + (-0.1 + 0.4)^2) / 2) = sqrt(0.07) = 0.26; 1.28*sigma = 0.33, so point +/- 0.33 gives about [-0.43, 0.23], rounded outward to an 80 percent interval of [-0.5, 0.3].

Counter-considerations: upside risk is a soft July CPI print below roughly 0.1 percent paired with another 0.3 percent nominal wage month, which would land above the interval. Downside risk is a renewed energy or shelter-driven CPI jump near 0.6 percent with only 0.2 percent wage growth, which would land below the interval. Outside the interval would most likely require either a large July CPI surprise or a meaningful payroll-composition shock to average hourly earnings.

Framing and exact resolver: this forecast is for BLS Real Earnings Table A-1, seasonally adjusted over-the-month percent change in real average hourly earnings for all employees on private nonfarm payrolls, first printed for July 2026. The variant is Table A-1 real average hourly earnings, all employees, private nonfarm, seasonally adjusted; all historical anchors below use that same variant.

Base rate/reference class: the official Table A-1 recent same-variant real AHE MoM prints available in the current table are 0.4, -0.6, -0.5, and -0.1 percent, averaging -0.2 percent, with only 1 of 4 positive. That base rate says the target is noisy but tilted slightly negative when CPI is running above normal.

Prior/update/interval: persistence prior is the same-variant Table A-1 recent-print mean of -0.2 percent from May 2025 and Mar-May 2026; update components are +0.1 for June nominal wage momentum staying near 0.3 percent, 0.0 for July labor-market softness not implying a wage break, and 0.0 for CPI likely still near nominal wage growth, giving point -0.1 percent. Interval method uses realized dispersion of the fetched change values [0.4, -0.6, -0.5, -0.1]: sample sigma = 0.46, half-width = 1.28*sigma = 1.28*0.46 = 0.59, rounded to the BLS one-decimal reporting grid around -0.1 gives an 80% interval of -0.7 to 0.5 percent.

Counter-considerations: upside risk is a benign July CPI print near 0.0 to 0.1 percent with nominal AHE still around 0.3 percent, which would land above the interval if real earnings printed at 0.6 percent or higher. Downside risk is another energy-driven CPI jump or a weak nominal AHE print, which would land below the interval if real earnings printed at -0.8 percent or lower. Outside the interval would most likely require a sharp one-month CPI or wage surprise rather than ordinary persistence.

Framing and exact resolver: this forecast is for BLS Real Earnings Table A-1, seasonally adjusted real average hourly earnings for all employees on private nonfarm payrolls, over-the-month percent change for July 2026, first print only, rounded to the one-decimal percent value printed by BLS.

Variant check: all real earnings anchors are the Table A-1 seasonally adjusted all-employees private-nonfarm real average hourly earnings series. FRED nominal AHE and CPI-U are used only to decompose the wage and deflator mechanisms, not as the resolution source.

Reference class and base rate: the recent official Table A-1 reference class has four fetched real AHE m/m prints, 0.4, -0.6, -0.5, and -0.1, with a mean base rate of -0.20 percent; the latest three prints average -0.40 percent, reflecting CPI running above nominal wage gains.

Prior/update/interval: persistence prior is the recent official Table A-1 mean from May 2025, Mar. 2026, Apr. 2026, and May 2026 real AHE m/m values: (0.4 - 0.6 - 0.5 - 0.1)/4 = -0.20. Adjustment components: +0.10 because nominal AHE momentum improved to about (37.64/37.51 - 1)*100 = 0.35 in June while July CPI is not assumed to repeat the March-May 0.5 to 0.9 pace; point = -0.10. Interval method: sample sigma from the four fetched real-change values is sigma = 0.45, so the 80 percent half-width is roughly 1.28*sigma = 0.58, rounded to 0.6; final implied bounds are -0.1 - 0.6 = -0.7 and -0.1 + 0.6 = 0.5.

Counter-consideration: upside risk is a soft July CPI print with another 0.3 to 0.4 nominal wage gain, which would land above the interval; downside risk is another CPI print near 0.7 or a weak July AHE print, which would land below the interval.

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 2026-07-10T16:27:44Z, 2026-07-10T16:39:16Z, 2026-07-10T16:50:39Z. Drivers and resolver fields mirror the rollout closest to the median.

Median CDF quantiles: q10 = -0.7, q50 = -0.1, q90 = 0.5. Constituent points [-0.1, -0.1, -0.1] with 80% widths [0.8, 1.2, 1.2]; the median interval inherits the central rollout mass rather than averaging tails.

The target is the first-print, one-decimal value in BLS Real Earnings Table A-1 for all employees on private nonfarm payrolls: seasonally adjusted real average hourly earnings over the month. Table A-1 uses CPI-U to deflate nominal earnings; subtracting CPI-U inflation from nominal wage growth is a close approximation rather than the exact index calculation. No later revision replaces the first print.

The outside-view base rate/reference class is the five documented first-print observations from January through May 2026: 0.3%, 0.2%, -0.6%, -0.5%, and -0.1%. Their mean is -0.14% and median is -0.1%. A formal time-series model is unsuitable because only five noisy first-print observations are documented and the target is a monthly change rather than a persistent level.

Mechanisms: typical nominal wage behavior is already embedded in the reference class, so it receives no separate adjustment. June nominal hourly earnings growth of 0.3% supports persistence but does not justify double-counting. The only directional adjustment is +0.1 percentage point from the prior median, reflecting partial normalization from the unusually high 0.5%-0.9% CPI-U readings associated with recent negative prints; absent direct July CPI evidence, the adjustment is deliberately small.

Prior/update/interval: Start from the five-print reference-class median of -0.1%. Using the change-series values themselves [0.3, 0.2, -0.6, -0.5, -0.1], the sample standard deviation is sigma = 0.40 percentage point, so 1.28*sigma = 0.52 point. Adjustment components are +0.10 point for partial CPI normalization and +0.00 for nominal-wage persistence already embedded in the prior, yielding a 0.0% center. The ladder-implied 80% interval is -0.5% to 0.5%, a 0.50-point half-width, close to the 0.52-point empirical width.

Ladder: P(X <= -0.6) = 0.05; P(X <= -0.5) = 0.10; P(X <= -0.4) = 0.17; P(X <= -0.3) = 0.25; P(X <= -0.2) = 0.34; P(X <= -0.1) = 0.43; P(X <= 0.0) = 0.50; P(X <= 0.1) = 0.59; P(X <= 0.2) = 0.68; P(X <= 0.3) = 0.77; P(X <= 0.4) = 0.84; P(X <= 0.5) = 0.90; P(X <= 0.6) = 0.95. Linear interpolation places the 10th, 50th, and 90th percentiles exactly at -0.5%, 0.0%, and 0.5%, respectively.

Upside risk comes from July CPI-U near zero while nominal hourly pay rises at least 0.4%; that could land above the interval. Downside risk comes from another energy- or tariff-related CPI jump near 0.8% with weak wage growth; that would land below the interval. Composition-driven hourly-pay volatility is an additional outside-the-interval mechanism.

Review disposition: Accepted all four warnings by removing the undocumented December observation, using the five cited prints, explaining why a formal time-series model is unsuitable, avoiding double-counted wage momentum, itemizing the CPI and wage adjustments, and shrinking the center to 0.0%. Also accepted the clarification that nominal growth minus CPI-U is approximate.

The target is BLS Real Earnings Table A-1: the seasonally adjusted over-the-month change in real average hourly earnings for all employees on private nonfarm payrolls. CPI-U is the deflator. Resolution uses the one-decimal July 2026 first print without later revisions.

The official BLS calendar schedules Real Earnings for July 2026 at 8:30 a.m. ET on August 12, 2026, verifying the ledger resolution date.

The reference class and base rate are the five available 2026 first-print changes: 0.3%, 0.2%, -0.6%, -0.5%, and -0.1%, with a mean of -0.14%. This series is a monthly change, so dispersion is measured from the values themselves.

Level, momentum, one-off, and policy mechanisms: nominal wage growth remains positive but slowed to 0.2% in June; recent real-earnings momentum is weak; unusually strong spring CPI prints produced much of that weakness; continuing price pressure can still offset ordinary wage growth in July.

Prior/update/interval: A mean/persistence prior from the five first-print 2026 observations is -0.14%. The historical sample is [0.3, 0.2, -0.6, -0.5, -0.1]. I adjust about +0.14 percentage point toward 0.0% because a roughly 0.3% nominal-pay gain and roughly 0.3% CPI-U gain would offset. Sample sigma = sqrt(0.652/4) = 0.40 percentage point; 1.28*sigma = 0.52, which rounds to a one-decimal-compatible half-width of 0.5. Final implied bounds are -0.5% to 0.5%.

Mechanism point estimate: forecast nominal average hourly earnings growth 0.3% minus forecast CPI-U inflation 0.3% = 0.0%. The 80% interval is 0.0% ± 1.28×0.40% = 0.0% ± 0.52%, reported as [-0.5%, 0.5%] in the target's one-decimal units.

Upside risk would come from a strong July wage print combined with softer CPI-U and could land above 0.5%. Downside risk would come from another tariff- or energy-driven inflation surge with subdued pay growth and could land below -0.5%; either outcome would be outside the interval.

The target is the first-print, seasonally adjusted monthly percent change for all private-nonfarm employees in BLS Real Earnings Table A-1. It is nominal average hourly earnings deflated by CPI-U; the first one-decimal July 2026 print is final for this forecast even if BLS later revises it.

The official BLS August 2026 release calendar verifies Real Earnings for July 2026 at 8:30 a.m. ET on August 12, 2026. The recent reference class/base rate is centered close to zero, with unusually negative March and April readings reflecting CPI-U increases of 0.9% and 0.6% against nominal wage gains of 0.2%.

Level, momentum, one-off, and mechanism: the real earnings level does not mechanically force mean reversion; recent momentum is negative; monthly energy and other CPI shocks can create one-off moves; mechanically, a roughly 0.3% nominal-pay gain minus roughly 0.3% CPI-U inflation implies about 0.0% real growth.

Prior/update/interval: A persistence/reference-class prior uses the ten available same-variant first-print observations June–September and December 2025 plus January–May 2026: -0.1, 0.1, -0.1, 0.0, 0.0, 0.3, 0.2, -0.6, -0.5, -0.1. Their mean is -0.08%; the nominal-pay adjustment is mildly positive, recent CPI momentum is negative, and no separate one-off adjustment is imposed, yielding 0.0%. For this change series, dispersion is measured from the values themselves: sample sigma = 0.28 percentage point, so 1.28*sigma = 0.36 point; rounding outward to the BLS one-decimal grid gives bounds of -0.4% and 0.4%.

Component estimate: 0.3% nominal hourly-pay growth - 0.3% CPI-U growth = 0.0% real growth. Historical 80% half-width = 1.28 × 0.28 = 0.36 percentage point; outward one-decimal bounds are 0.0 - 0.4 = -0.4% and 0.0 + 0.4 = 0.4%.

Upside risk: softer energy or goods inflation alongside a 0.4% wage print would land above the interval. Downside risk: another CPI-U surge near 0.8–0.9% with only 0.2% nominal wage growth would land below the interval. Either outcome would be outside the interval and falsify the central mechanism assumptions.

The target is the first-print, seasonally adjusted month-over-month percentage change for all employees on private nonfarm payrolls in BLS Real Earnings Table A-1. The all-employees series uses CPI-U as its deflator; production and nonsupervisory earnings and CPI-W are different variants and are excluded.

The recent first-print reference class is December 2025 through May 2026: -0.3, 0.3, 0.2, -0.6, -0.5, and -0.1 percent. Its base rate mean is -0.17 percent, reflecting several recent months when CPI-U increases exceeded nominal wage gains.

Mechanisms: the level of real hourly earnings was $11.24 in May; nominal wage momentum has recently been about 0.2–0.4 percent monthly; CPI-U is the principal offset; and volatile monthly inflation is the dominant one-off risk. For July, a roughly 0.3 percent nominal-pay gain and roughly 0.3 percent CPI-U increase imply approximately no real change.

Prior/update/interval: The persistence/reference-class prior is the -0.17 percent mean of the six first-print changes [-0.3, 0.3, 0.2, -0.6, -0.5, -0.1]. For this change series, dispersion is computed from the values themselves: sample sigma = 0.37 percentage point. The normal 80% half-width is 1.28*sigma = 1.28*0.37 = 0.47 point. Updating the prior by about +0.17 point for an expected normalization of CPI-U relative to March–May, while retaining ordinary nominal-pay momentum, gives a 0.0 point estimate; rounding the bounds to the BLS one-decimal grid gives -0.5 to 0.5 percent.

Counter-considerations: upside risk would come from a soft or falling July CPI-U print alongside firm nominal wages and could put real earnings above 0.5 percent. Downside risk would come from another inflation surge or weak payroll earnings and could put the result below -0.5 percent. Either outcome would land outside the interval.

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 2026-07-10T17:21:35Z, 2026-07-10T17:27:05Z, 2026-07-10T17:32:35Z. Drivers and resolver fields mirror the rollout closest to the median.

Median CDF quantiles: q10 = -0.5, q50 = 0.0, q90 = 0.5. Constituent points [0.0, 0.0, 0.0] with 80% widths [1.0, 0.8, 1.0]; the median interval inherits the central rollout mass rather than averaging tails.

Framing and exact resolver: this forecast is for BLS Real Earnings Table A-1, all employees on private nonfarm payrolls, seasonally adjusted, real average hourly earnings over-the-month percent change for July 2026. The target resolves to the one-decimal first print on the BLS official release page, with no later revision used.

The same-variant rule matters: all anchors above refer either directly to BLS Table A-1 real average hourly earnings for all employees on private nonfarm payrolls, seasonally adjusted, or to its BLS-published nominal AHE and CPI-U inputs. I am not using FRED or local catalog estimates as evidence.

Reference class and base rate: recent Table A-1 real hourly MoM prints centered below zero, with Mar.-May 2026 at -0.6, -0.5, and -0.1 after CPI-U rose 0.9, 0.6, and 0.5. A simple AR or persistence model and a short historical mean both point negative; I use persistence rather than a fitted model because this target is a one-month first-print combination of wages and CPI, and the available same-variant recent sample is short.

Prior/update/interval: starting from a persistence prior near the recent Table A-1 average of about -0.4 for Mar.-May 2026, I add about +0.2 because June nominal AHE momentum was still roughly 0.3 percent, add about +0.1 because May core CPI was only 0.2 despite headline energy pressure, and subtract a small residual risk adjustment for headline CPI and energy volatility visible in the Mar.-May CPI prints. June AHE and May CPI are momentum indicators, not direct July target inputs. The threshold span is a subjective 10th-90th elicitation anchored by fetched same-variant real AHE outcomes of -0.6, -0.5, -0.1, and +0.4 rather than a stable volatility estimate, so [-0.5, 0.3] covers a repeat negative month through a modest rebound while leaving larger CPI or wage surprises in the tails.

Ladder: P(X <= -0.9) = 0.02; P(X <= -0.7) = 0.06; P(X <= -0.5) = 0.10; P(X <= -0.4) = 0.17; P(X <= -0.3) = 0.26; P(X <= -0.2) = 0.38; P(X <= -0.1) = 0.52; P(X <= 0.0) = 0.65; P(X <= 0.1) = 0.76; P(X <= 0.2) = 0.84; P(X <= 0.3) = 0.90; P(X <= 0.5) = 0.97; P(X <= 0.7) = 0.99. Linear interpolation gives 10th percentile at -0.5, median at -0.1142857142857143, and 90th percentile at 0.3; rounded to BLS one-decimal print precision, the forecast is point -0.1 with 80 percent interval [-0.5, 0.3].

Upside risk: if July CPI-U cools sharply while nominal AHE keeps a 0.3 percent or better monthly pace, the real hourly print would land above the interval. Downside risk: renewed energy-price pressure, CPI above roughly 0.5 percent, or July nominal AHE below roughly 0.1 percent could put the deflator well above nominal wage growth and would land below the interval. An outside the interval outcome is most likely from a large CPI surprise or an unusually large July payroll wage surprise.

Review disposition: accepted the interval critique by stating the ladder is a subjective elicitation anchored to the short same-variant reference class rather than a fitted volatility estimate; accepted the update critique by clarifying June AHE and May CPI are only momentum indicators and softening the energy adjustment; accepted the model-prior critique by naming the simple AR/persistence prior and why it is sufficient here.

The target is the first printed one-decimal July 2026 over-the-month change in seasonally adjusted real average hourly earnings for all employees on private nonfarm payrolls in BLS Table A-1. Table A-1 uses CPI-U to deflate the all-employees CES earnings series; later revisions do not alter resolution.

This limited, non-contiguous reference class has a base rate near zero but substantial monthly variation: the fetched same-series prints span -0.6% to 0.3%, and the July 2025 analogue was 0.1%. Recent 2026 momentum is weaker, with March through May at -0.6%, -0.5%, and -0.1%.

Mechanism decomposition: nominal hourly-pay growth is typically positive and comparatively stable, while monthly CPI-U is the volatile subtraction. I anchor nominal earnings near 0.3% and allow CPI-U near 0.3%-0.4%, producing a slightly negative real change. No separate hours adjustment enters this hourly target.

Prior/update/interval: The persistence prior is the latest fetched first print, -0.1%; the median of the limited five-observation reference class is also -0.1%. I apply a -0.1 percentage-point momentum update for the three latest 2026 prints being negative, then a +0.1 percentage-point normalization update because May's 0.5% CPI-U increase was unusually strong relative to its 0.3% nominal-pay increase. These updates net to zero and retain a -0.1% point estimate. A separate formal time-series model was ruled out because the fetched first-print sample is limited and non-contiguous. The -0.6% and 0.3% fetched extremes anchor the ladder span, and linear interpolation of the ladder gives rounded 80% bounds of -0.5% and 0.3%.

Upside risk comes from softer-than-expected July CPI-U or a strong nominal-pay print and could put real earnings above 0.3%. Downside risk comes from another inflation surge combined with soft wage growth and could put the result below -0.5%; either outcome would land outside the interval.

Ladder: P(X <= -0.8) = 0.02; P(X <= -0.6) = 0.05; P(X <= -0.5) = 0.09; P(X <= -0.4) = 0.15; P(X <= -0.3) = 0.25; P(X <= -0.2) = 0.38; P(X <= -0.1) = 0.52; P(X <= 0) = 0.65; P(X <= 0.1) = 0.76; P(X <= 0.2) = 0.85; P(X <= 0.3) = 0.91; P(X <= 0.4) = 0.95; P(X <= 0.6) = 0.98. Linear interpolation gives the 10th percentile at -0.483%, median at -0.114%, and 90th percentile at 0.283%; at BLS one-decimal print precision these are -0.5%, -0.1%, and 0.3%.

Review disposition: Accepted the request to quantify the prior and adjustment components, state how they yield the point estimate, disclose that no formal time-series model was used, and label the historical sample as limited and non-contiguous.

The resolver is BLS Real Earnings Table A-1: the seasonally adjusted over-the-month percent-change row for real average hourly earnings of all employees on private nonfarm payrolls. Every anchor uses this exact variant; Table A-1 deflates nominal average hourly earnings using CPI-U.

The reference class/base rate is the January 2016-February 2026 sequence of monthly seasonally adjusted Table A-1 real-hourly-earnings changes. I do not fit a separate predictive regression because the target is a one-decimal, short-horizon release with inflation, nominal-wage, and composition shocks that are not reliably known before release; the historical median near +0.1 percent and central 80 percent range of about -0.7 to +0.8 percent are the reproducible prior.

The March-to-May sequence (-0.6, -0.5, -0.1) updates the +0.1 percent historical-median prior downward, but the narrowing May shortfall argues against extending the earlier declines mechanically. Nominal earnings growth relative to CPI-U is the main near-term mechanism; changing private-nonfarm composition and seasonal adjustment remain residual sources of variation.

Prior/update/interval: The historical prior is the January 2016-February 2026 Table A-1 reference class (122 monthly prints), with median about +0.1 percent and central 80 percent about -0.7 to +0.8 percent. The current-evidence update is the fetched March -0.6, April -0.5, and May -0.1 sequence, plus May's +0.3 percent nominal earnings change against +0.5 percent CPI-U, which moves the center to 0.0 percent while retaining roughly historical realized monthly dispersion for wage-growth, CPI-U, composition, and seasonal uncertainty. Ladder: P(X <= -1.0) = 0.03; P(X <= -0.8) = 0.06; P(X <= -0.6) = 0.13; P(X <= -0.4) = 0.23; P(X <= -0.2) = 0.37; P(X <= 0.0) = 0.53; P(X <= 0.2) = 0.68; P(X <= 0.4) = 0.79; P(X <= 0.6) = 0.87; P(X <= 0.8) = 0.93; P(X <= 1.0) = 0.96; P(X <= 1.2) = 0.98; P(X <= 1.4) = 0.99. Linear interpolation gives 10th percentile at -0.7, median at 0.0, and 90th percentile at 0.7; rounded to BLS's one-decimal print precision, the point is 0.0 percent and the 80% interval is -0.7 to 0.7 percent.

upside risk: a soft or negative July CPI-U print combined with firm nominal earnings could lift real hourly earnings above 0.7 percent. downside risk: a renewed CPI-U jump while nominal earnings growth slows could push the print below -0.7 percent. A larger energy-price or composition shock would land outside the interval.

Review disposition: Accepted the requested ordering, explicit January 2016-February 2026 reference-class prior, quantified historical interval basis, and separate current-evidence update; retained the ladder-derived forecast because its bounds remain coherent with that evidence.

The target is the BLS CES series bls.real_earnings.avg_hourly_mom, specifically Table A-1 for all employees on private nonfarm payrolls, seasonally adjusted, and resolved at the first official July 2026 print without later revisions.

The base rate/reference class is the five fetched 2026 observations: 0.3, 0.1, -0.6, -0.5, and -0.1 percent. Their median is -0.1 percent, with four of five values between -0.6 and 0.3.

The explicit model prior is a persistence time-series prior centered on the recent five-observation median, -0.1 percent. I do not use a richer autoregressive or structural model because the fetched sample is short and monthly real-earnings prints contain substantial deflator and seasonal-adjustment noise.

The exact variant is the gross all-employee, private nonfarm payrolls, seasonally adjusted real average hourly earnings series in BLS Table A-1, deflated with CPI-U; no production-and-nonsupervisory, NSA, weekly-earnings, or revised-vintage values are mixed in.

Prior/update/interval: The persistence model prior is -0.1 percent. Recent negative momentum and CPI-U volatility are acknowledged, but no dated current evidence in the fetched release history warrants moving the point estimate from that prior. The fetched reference class anchors the ladder span from -0.6 to 0.3 percent; the ladder's interpolated 10th and 90th percentiles produce the 80% interval from -0.7 to 0.5 percent.

Ladder: P(X <= -1) = 0.05; P(X <= -0.7) = 0.1; P(X <= -0.5) = 0.2; P(X <= -0.3) = 0.32; P(X <= -0.1) = 0.46; P(X <= 0) = 0.56; P(X <= 0.1) = 0.68; P(X <= 0.3) = 0.8; P(X <= 0.5) = 0.89; P(X <= 0.7) = 0.95; P(X <= 1) = 0.98. Linear interpolation gives 10th percentile at -0.7, median at -0.06, and 90th percentile at 0.53; rounding to BLS one-decimal print precision gives 10th percentile at -0.7, median at -0.1, and 90th percentile at 0.5.

Downside risk is renewed CPI acceleration with nominal pay below trend, which would land below the interval; upside risk is softer CPI alongside firm nominal wage growth, which would land above the interval. An unusually large seasonal-adjustment change or release-month data shock would be outside the interval.

Review disposition: Accepted the critique by explicitly naming the persistence time-series prior, explaining why no richer model is used, and stating that current fetched evidence does not warrant moving the point from the prior; clarified that the bounds are ladder-interpolated percentiles.