thesisPrototype

Government data

UK regular pay growth, May-Jul 2026

ONS KAI9 AWE whole economy year-on-year three-month average growth, seasonally adjusted regular pay excluding arrears, value for 2026 JUL, first print

Overdue · was due Sep 15, 2026

The resolver has no route for this target's reference. The target was registered with a generic source link rather than a resolver binding.

NO_ROUTE_GENERIC_URL · resolver status as of Oct 7, 2026

Forecast

3.5%
3.1%–3.9%
80% prediction interval
gpt-5.5Run record ↗
2.83.64.35.1July 2025June 2026Sep 20263.5%
historyforecast path80% interval
Probability distribution

Cumulative probability

Chance that the outcome is at or below each value.

0%25%50%75%100%2.5%3.5%4.5%Forecast value

Shaded band: 80% interval (3.1%–3.9%). Dashed line: point estimate (3.5%).

Derived from the point estimate and 80% interval; the agent did not report a full distribution.

Analysis

§UK regular pay growth first-print forecast

Framing and exact resolver: the target is ONS series KAI9, AWE whole economy year-on-year three-month average growth (%), seasonally adjusted regular pay excluding arrears, value for 2026 JUL. The official ONS UK Labour Market: September 2026 page verifies the release date as 15 September 2026 at 7:00am; I use the first KAI9 print and ignore later revisions.

Variant discipline: all anchors below are from the same KAI9 variant named in the resolver: regular pay excluding arrears, seasonally adjusted, whole economy, year-on-year three-month average growth in percent.

▸ Reported tool use: official.lookupmodel report
ONS KAI9 time-series page, latest observed target-variant prints
↳ Fetched KAI9 recent values: 2026 JAN 3.8, 2026 FEB 3.6, 2026 MAR 3.4, 2026 APR 3.4, 2026 MAY 3.4, 2026 JUN 3.5 percent.
▸ Reported tool use: official.lookupmodel report
ONS KAI9 time-series page, prior six monthly prints for the same variant
↳ Fetched KAI9 prior values: 2025 JUL 4.8, 2025 AUG 4.7, 2025 SEP 4.7, 2025 OCT 4.6, 2025 NOV 4.4, 2025 DEC 4.1 percent.
▸ Reported tool use: official.lookupmodel report
ONS KAI9 metadata and ONS September 2026 labour-market release page
↳ Fetched metadata and schedule: KAI9 release date 18 August 2026, next release 15 September 2026, latest observed KAI9 value 2026 JUN 3.5, and UK Labour Market: September 2026 release time 7:00am.

Base rate/reference class: the 12 fetched KAI9 prints from 2025 JUL through 2026 JUN average 4.03 percent, but the last six average 3.52 percent and the last four sit at 3.4, 3.4, 3.4, and 3.5, so persistence around 3.5 is the base rate for a one-month-ahead three-month-average target.

Prior/update/interval: persistence prior is latest KAI9 = 3.5 using the historical sample 2025 JUL-2026 JUN. Level adjustment is 0.0 because the latest six-print mean is 3.52 and the latest print is 3.5; momentum adjustment is 0.0 because the series flattened at 3.4-3.5; one-off adjustment is 0.0 because public-sector pay-award support is offset by private-sector cooling; policy-mechanism adjustment is 0.0 because this is a labour-market survey print, not a policy decision. For one-month-ahead uncertainty I use realized monthly changes in the fetched KAI9 prints: [-0.1, 0.0, -0.1, -0.2, -0.3, -0.3, -0.2, -0.2, 0.0, 0.0, 0.1], giving sigma = 0.133 and 1.28*sigma = 0.170. I widen to a 0.4 point half-width because the short realized-change sample is a smooth disinflation plateau and the next three-month average replaces one component month with July; rounded to one decimal, the 80% interval is 3.1 to 3.9.

Upside risk: a stronger July single-month pay reading, renewed public-sector award effects, or composition shifts toward higher-paid employment would land above the interval if they push KAI9 above 3.9. Downside risk: sharper private-sector wage cooling, weaker paid-hours composition, or survey reweighting would land below the interval if they push KAI9 below 3.1. Outside the interval would require a break from the recent 3.4-3.5 plateau rather than ordinary month-to-month noise.

Review disposition: accepted the interval critique by switching the uncertainty arithmetic from level dispersion to realized monthly changes and narrowing the interval; accepted the suggestion to make the one-off adjustment a single 0.0 value and kept source context focused on the exact KAI9 and September 2026 ONS pages.

Key drivers

  • latest KAI9 print was 3.5 percent
  • recent six-print plateau near 3.5 percent
  • private-sector wage cooling restrains upside
  • public-sector pay awards support the aggregate
  • three-month averaging smooths one-month July movement

Tool evidence

This run has no captured tool responses. “Reported tool use” in the analysis is the model’s account, not an archived tool response.

Sources and resolution

Official source
Office for National Statistics Labour market statistics time series (LMS), KAI9
Resolution date
September 15, 2026· outcome not recorded
Resolution rule
Resolve to the first ONS-published KAI9 value for 2026 JUL in the Labour market statistics time series: AWE whole economy year-on-year three-month average growth (%), seasonally adjusted regular pay excluding arrears. Use the first print published for UK Labour Market: September 2026 on 15 September 2026, as displayed on the KAI9 time-series page, in percent and rounded as ONS publishes it; ignore later revisions or historical restatements.
Run details

The analysis is the model’s written report. Tool-use descriptions in that report are model claims; the activity artifacts contain the execution record.

thesis.analyst · gpt-5.5 · fast · v2.5.11

pre-submit review · completed

Draft is publishable with one methodological warning: the resolver and JSON fields align with the ledger, but the interval basis should be tightened.

  • warning interval: The 80% interval is sized from the cross-sectional dispersion of the last 12 level values, not from month-to-month realized volatility, first-print forecast errors, or another explicit uncertainty model for a one-month-ahead three-month-average target.
  • info optional_suggestion: Make the compact Prior/update/interval step state the final one-off adjustment as a single value rather than '+0.0 to +0.1' since the point estimate applies no visible adjustment.
  • info optional_suggestion: Consider replacing the generic 'latest' bulletin citation with the exact observed KAI9 page/release evidence already named, to reduce link-rot ambiguity.

disposition accepted: Review disposition: accepted the interval critique by switching the uncertainty arithmetic from level dispersion to realized monthly changes and narrowing the interval; accepted the suggestion to make the one-off adjustment a single 0.0 value and kept source context focused on the exact KAI9 and September 2026 ONS pages.

disposition not applicable: Review disposition: accepted the interval critique by switching the uncertainty arithmetic from level dispersion to realized monthly changes and narrowing the interval; accepted the suggestion to make the one-off adjustment a single 0.0 value and kept source context focused on the exact KAI9 and September 2026 ONS pages.

disposition not applicable: Review disposition: accepted the interval critique by switching the uncertainty arithmetic from level dispersion to realized monthly changes and narrowing the interval; accepted the suggestion to make the one-off adjustment a single 0.0 value and kept source context focused on the exact KAI9 and September 2026 ONS pages.

Activity artifacts

Complete original trace

§UK regular pay growth first-print forecast

Framing and exact resolver: the target is ONS series KAI9, AWE whole economy year-on-year three-month average growth (%), seasonally adjusted regular pay excluding arrears, value for 2026 JUL. The official ONS UK Labour Market: September 2026 page verifies the release date as 15 September 2026 at 7:00am; I use the first KAI9 print and ignore later revisions.

Variant discipline: all anchors below are from the same KAI9 variant named in the resolver: regular pay excluding arrears, seasonally adjusted, whole economy, year-on-year three-month average growth in percent.

▸ Reported tool use: official.lookupmodel report
ONS KAI9 time-series page, latest observed target-variant prints
↳ Fetched KAI9 recent values: 2026 JAN 3.8, 2026 FEB 3.6, 2026 MAR 3.4, 2026 APR 3.4, 2026 MAY 3.4, 2026 JUN 3.5 percent.
▸ Reported tool use: official.lookupmodel report
ONS KAI9 time-series page, prior six monthly prints for the same variant
↳ Fetched KAI9 prior values: 2025 JUL 4.8, 2025 AUG 4.7, 2025 SEP 4.7, 2025 OCT 4.6, 2025 NOV 4.4, 2025 DEC 4.1 percent.
▸ Reported tool use: official.lookupmodel report
ONS KAI9 metadata and ONS September 2026 labour-market release page
↳ Fetched metadata and schedule: KAI9 release date 18 August 2026, next release 15 September 2026, latest observed KAI9 value 2026 JUN 3.5, and UK Labour Market: September 2026 release time 7:00am.

Base rate/reference class: the 12 fetched KAI9 prints from 2025 JUL through 2026 JUN average 4.03 percent, but the last six average 3.52 percent and the last four sit at 3.4, 3.4, 3.4, and 3.5, so persistence around 3.5 is the base rate for a one-month-ahead three-month-average target.

Prior/update/interval: persistence prior is latest KAI9 = 3.5 using the historical sample 2025 JUL-2026 JUN. Level adjustment is 0.0 because the latest six-print mean is 3.52 and the latest print is 3.5; momentum adjustment is 0.0 because the series flattened at 3.4-3.5; one-off adjustment is 0.0 because public-sector pay-award support is offset by private-sector cooling; policy-mechanism adjustment is 0.0 because this is a labour-market survey print, not a policy decision. For one-month-ahead uncertainty I use realized monthly changes in the fetched KAI9 prints: [-0.1, 0.0, -0.1, -0.2, -0.3, -0.3, -0.2, -0.2, 0.0, 0.0, 0.1], giving sigma = 0.133 and 1.28*sigma = 0.170. I widen to a 0.4 point half-width because the short realized-change sample is a smooth disinflation plateau and the next three-month average replaces one component month with July; rounded to one decimal, the 80% interval is 3.1 to 3.9.

Upside risk: a stronger July single-month pay reading, renewed public-sector award effects, or composition shifts toward higher-paid employment would land above the interval if they push KAI9 above 3.9. Downside risk: sharper private-sector wage cooling, weaker paid-hours composition, or survey reweighting would land below the interval if they push KAI9 below 3.1. Outside the interval would require a break from the recent 3.4-3.5 plateau rather than ordinary month-to-month noise.

Review disposition: accepted the interval critique by switching the uncertainty arithmetic from level dispersion to realized monthly changes and narrowing the interval; accepted the suggestion to make the one-off adjustment a single 0.0 value and kept source context focused on the exact KAI9 and September 2026 ONS pages.

calibrated forecast · 80% CI
3.5%[3.1% · 3.9%]
Target metadata

Data point: ons.earnings.regular_pay_yoy.2026_07.first_print

More government data forecasts