thesisPrototype

Government data

Belgium NBB business barometer, August 2026

National Bank of Belgium monthly business surveys, overall synthetic curve for Belgium, total sector, seasonally adjusted, August 2026, first official print

Overdue · was due Aug 28, 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

-11.9
-15.1–-8.7
80% prediction interval
gpt-5.5Run record ↗

Cumulative probability

Chance that the outcome is at or below each value.

0%25%50%75%100%-19.9-11.9-3.9Forecast value

Shaded band: 80% interval (-15.1–-8.7). Dashed line: point estimate (-11.9).

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

Analysis

§Belgium August 2026 NBB Business Barometer Forecast

Framing and resolver: the registered target is nbb.business_barometer.overall for 2026-08, country BE, unit index_points, dataPointId nbb.business_barometer.overall.2026-08.first_print, and ledger resolutionDate 2026-08-28. The exact resolving series is the National Bank of Belgium monthly business surveys overall synthetic curve, Belgium, total sector, seasonally adjusted, first print. I fetched https://app.thesisinstitute.org/specs.json during the run; it returned a 404 HTML page rather than a usable specs JSON, so I kept the registered ledger slug belgium-nbb-business-barometer-august-2026.

▸ Reported tool use: nbb.dsdmodel report
curl -L -sS 'https://nsidisseminate-stat.nbb.be/rest/datastructure/BE2/DSD_BUSSURVM/1.0?references=none'
↳ Fetched NBB DSD at prepared time 2026-08-13T07:17:04Z. Dimension order is 1 FREQ, 2 BUSSURVM_INDICATOR, 3 BE_AREA, 4 BUSSURV_SECTOR, 5 BUSSURVM_ADJ, 6 TIME_PERIOD; target key is M.SYNC.BE.A999.S, with DECIMALS=1 for one-decimal index-point observations.
▸ Reported tool use: nbb.sdmxmodel report
curl -L -sS -H 'Accept: application/vnd.sdmx.data+json;version=1.0.0-wd' 'https://nsidisseminate-stat.nbb.be/rest/data/DF_BUSSURVM/M.SYNC.BE.A999.S?startPeriod=2025-08&endPeriod=2026-07&dimensionAtObservation=TIME_PERIOD'
↳ Fetched official NBB overall seasonally adjusted synthetic curve values: 2025-08 -8.9, 2025-09 -7.9, 2025-10 -9.1, 2025-11 -8.2, 2025-12 -11.9, 2026-01 -8.8, 2026-02 -13.7, 2026-03 -13.7, 2026-04 -14.2, 2026-05 -13.3, 2026-06 -12.4, 2026-07 -11.9 index points.
▸ Reported tool use: nbb.sdmxmodel report
curl -L -sS -H 'Accept: application/vnd.sdmx.data+json;version=1.0.0-wd' 'https://nsidisseminate-stat.nbb.be/rest/data/DF_BUSSURVM/M.SYNC.BE.M000+T000+S000.S?startPeriod=2026-06&endPeriod=2026-07&dimensionAtObservation=TIME_PERIOD'
↳ Fetched official NBB sector context: manufacturing industry 2026-06 -15.7 and 2026-07 -15.5; trade 2026-06 -9.3 and 2026-07 -10.8; business-related services 2026-06 -7.4 and 2026-07 -3.8 index points.
▸ Reported tool use: model.candidatesmodel report
Computed thesis_model_candidate_v1 candidates from the fetched NBB 2025-08 through 2026-07 total synthetic-curve history.
↳ thesis_model_candidate_v1 persistence candidate: point -11.9, p10 -15.1, p50 -11.9, p90 -8.7, interval80 [-15.1,-8.7], interval90 [-15.6,-8.2], intervalMethod residual_change_sigma_widened, calibration_n 11, trainCutoff 2026-07, walkForwardMae 1.60. trailing3_mean candidate: point -12.5, p10 -15.7, p50 -12.5, p90 -9.3, interval80 [-15.7,-9.3], interval90 [-16.2,-8.8], intervalMethod residual_change_sigma_widened, calibration_n 9, trainCutoff 2026-07, walkForwardMae 1.84.

Base rate/reference class: trailing 12 official NBB total synthetic-curve prints range from -14.2 to -7.9 and average -11.2; the latest six average -13.2, the latest three average -12.5, and the latest print is -11.9. Because persistence beats the trailing-3 mean in the walk-forward check, the outside-view base rate and strongest benchmark is last-print persistence at -11.9.

Inside-view update: July had a third consecutive improvement from the April trough, and business-related services improved from -7.4 to -3.8, but trade worsened from -9.3 to -10.8 and manufacturing was nearly flat at -15.5. With no direct August survey signal before the print, I do not move materially away from the persistence benchmark.

Prior/update/interval: persistence prior = -11.9 from the July 2026 official print; historical sample = NBB total synthetic curve 2025-08 through 2026-07 with changes +1.0,-1.2,+0.9,-3.7,+3.1,-4.9,+0.0,-0.5,+0.9,+0.9,+0.5. Adjustment components: +0.2 for recent improvement momentum, -0.2 for mixed sector composition and no direct August signal, net 0.0, so point = -11.9. For the interval, sigma = 2.28 from successive fetched monthly changes; 1.28*sigma = 2.91, widened modestly to half-width 3.2 because only 7 of the last 10 changes fit within 2.91 while 8 of 10 fit within 3.2. Final 80% bounds: -11.9 - 3.2 = -15.1 and -11.9 + 3.2 = -8.7.

Counter-consideration and falsification: upside risk is a broad August improvement in manufacturing and trade matching July's services rebound, which would land above the interval if the first print is -8.6 or higher. Downside risk is renewed trade/export weakness or cost-pressure pessimism spreading across sectors, which would land below the interval if the first print is -15.2 or lower. Either outside the interval would be a sharper one-month move than the widened recent-change reference class expects.

Review disposition: accepted the required traceability fix by aligning the model-candidate sample, monthly-change list, calibration counts, sigma, and interval explanation to the explicitly fetched 2025-08 through 2026-07 NBB overall-series data. Accepted the resolver-clarity suggestion by stating that the registered NBB business survey page remains the resolving URL while SDMX was used for evidence.

Key drivers

  • July first print was -11.9 after -12.4 in June
  • persistence beats a trailing-3 mean in the walk-forward check
  • services improved in July while trade weakened
  • no direct August survey signal is available before the print
  • monthly first-print changes remain volatile

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
National Bank of Belgium business survey release
Resolution date
August 28, 2026· outcome not recorded
Resolution rule
Resolve to the first National Bank of Belgium published monthly business surveys overall synthetic curve for Belgium, total sector, seasonally adjusted, period 2026-08, reported in index points and rounded to one decimal. SDMX identity: DF_BUSSURVM key M.SYNC.BE.A999.S (FREQ monthly, BUSSURVM_INDICATOR synthetic curve, BE_AREA Belgium, BUSSURV_SECTOR Total, BUSSURVM_ADJ seasonally adjusted). Use the first official print only; ignore later NBB.Stat revisions, corrections after the first release, and third-party mirrors. The resolving URL remains the registered NBB business survey page; the NBB SDMX API was used only for pre-resolution history and series-identity evidence.
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 · full · v2.5.9

pre-submit review · completed

The draft is publishable after tightening one traceability issue around the historical sample used for the model and interval.

  • warning prior_update_interval: The model candidate and interval step say they use NBB history from 2025-07 through 2026-07, but the cited overall-series fetch in the trace starts at 2025-08 and does not show the 2025-07 observation needed for the first listed monthly change.
  • info optional_suggestion: Remove or de-emphasize the failed specs.json fetch unless it is needed for audit context.
  • info optional_suggestion: State explicitly that the resolving URL remains the registered NBB business survey page even though the evidence trace also used the NBB SDMX API.

disposition accepted: Review disposition: accepted the required traceability fix by aligning the model-candidate sample, monthly-change list, calibration counts, sigma, and interval explanation to the explicitly fetched 2025-08 through 2026-07 NBB overall-series data. Accepted the resolver-clarity suggestion by stating that the registered NBB business survey page remains the resolving URL while SDMX was used for evidence.

disposition not applicable: Review disposition: accepted the required traceability fix by aligning the model-candidate sample, monthly-change list, calibration counts, sigma, and interval explanation to the explicitly fetched 2025-08 through 2026-07 NBB overall-series data. Accepted the resolver-clarity suggestion by stating that the registered NBB business survey page remains the resolving URL while SDMX was used for evidence.

disposition not applicable: Review disposition: accepted the required traceability fix by aligning the model-candidate sample, monthly-change list, calibration counts, sigma, and interval explanation to the explicitly fetched 2025-08 through 2026-07 NBB overall-series data. Accepted the resolver-clarity suggestion by stating that the registered NBB business survey page remains the resolving URL while SDMX was used for evidence.

Activity artifacts

Complete original trace

§Belgium August 2026 NBB Business Barometer Forecast

Framing and resolver: the registered target is nbb.business_barometer.overall for 2026-08, country BE, unit index_points, dataPointId nbb.business_barometer.overall.2026-08.first_print, and ledger resolutionDate 2026-08-28. The exact resolving series is the National Bank of Belgium monthly business surveys overall synthetic curve, Belgium, total sector, seasonally adjusted, first print. I fetched https://app.thesisinstitute.org/specs.json during the run; it returned a 404 HTML page rather than a usable specs JSON, so I kept the registered ledger slug belgium-nbb-business-barometer-august-2026.

▸ Reported tool use: nbb.dsdmodel report
curl -L -sS 'https://nsidisseminate-stat.nbb.be/rest/datastructure/BE2/DSD_BUSSURVM/1.0?references=none'
↳ Fetched NBB DSD at prepared time 2026-08-13T07:17:04Z. Dimension order is 1 FREQ, 2 BUSSURVM_INDICATOR, 3 BE_AREA, 4 BUSSURV_SECTOR, 5 BUSSURVM_ADJ, 6 TIME_PERIOD; target key is M.SYNC.BE.A999.S, with DECIMALS=1 for one-decimal index-point observations.
▸ Reported tool use: nbb.sdmxmodel report
curl -L -sS -H 'Accept: application/vnd.sdmx.data+json;version=1.0.0-wd' 'https://nsidisseminate-stat.nbb.be/rest/data/DF_BUSSURVM/M.SYNC.BE.A999.S?startPeriod=2025-08&endPeriod=2026-07&dimensionAtObservation=TIME_PERIOD'
↳ Fetched official NBB overall seasonally adjusted synthetic curve values: 2025-08 -8.9, 2025-09 -7.9, 2025-10 -9.1, 2025-11 -8.2, 2025-12 -11.9, 2026-01 -8.8, 2026-02 -13.7, 2026-03 -13.7, 2026-04 -14.2, 2026-05 -13.3, 2026-06 -12.4, 2026-07 -11.9 index points.
▸ Reported tool use: nbb.sdmxmodel report
curl -L -sS -H 'Accept: application/vnd.sdmx.data+json;version=1.0.0-wd' 'https://nsidisseminate-stat.nbb.be/rest/data/DF_BUSSURVM/M.SYNC.BE.M000+T000+S000.S?startPeriod=2026-06&endPeriod=2026-07&dimensionAtObservation=TIME_PERIOD'
↳ Fetched official NBB sector context: manufacturing industry 2026-06 -15.7 and 2026-07 -15.5; trade 2026-06 -9.3 and 2026-07 -10.8; business-related services 2026-06 -7.4 and 2026-07 -3.8 index points.
▸ Reported tool use: model.candidatesmodel report
Computed thesis_model_candidate_v1 candidates from the fetched NBB 2025-08 through 2026-07 total synthetic-curve history.
↳ thesis_model_candidate_v1 persistence candidate: point -11.9, p10 -15.1, p50 -11.9, p90 -8.7, interval80 [-15.1,-8.7], interval90 [-15.6,-8.2], intervalMethod residual_change_sigma_widened, calibration_n 11, trainCutoff 2026-07, walkForwardMae 1.60. trailing3_mean candidate: point -12.5, p10 -15.7, p50 -12.5, p90 -9.3, interval80 [-15.7,-9.3], interval90 [-16.2,-8.8], intervalMethod residual_change_sigma_widened, calibration_n 9, trainCutoff 2026-07, walkForwardMae 1.84.

Base rate/reference class: trailing 12 official NBB total synthetic-curve prints range from -14.2 to -7.9 and average -11.2; the latest six average -13.2, the latest three average -12.5, and the latest print is -11.9. Because persistence beats the trailing-3 mean in the walk-forward check, the outside-view base rate and strongest benchmark is last-print persistence at -11.9.

Inside-view update: July had a third consecutive improvement from the April trough, and business-related services improved from -7.4 to -3.8, but trade worsened from -9.3 to -10.8 and manufacturing was nearly flat at -15.5. With no direct August survey signal before the print, I do not move materially away from the persistence benchmark.

Prior/update/interval: persistence prior = -11.9 from the July 2026 official print; historical sample = NBB total synthetic curve 2025-08 through 2026-07 with changes +1.0,-1.2,+0.9,-3.7,+3.1,-4.9,+0.0,-0.5,+0.9,+0.9,+0.5. Adjustment components: +0.2 for recent improvement momentum, -0.2 for mixed sector composition and no direct August signal, net 0.0, so point = -11.9. For the interval, sigma = 2.28 from successive fetched monthly changes; 1.28*sigma = 2.91, widened modestly to half-width 3.2 because only 7 of the last 10 changes fit within 2.91 while 8 of 10 fit within 3.2. Final 80% bounds: -11.9 - 3.2 = -15.1 and -11.9 + 3.2 = -8.7.

Counter-consideration and falsification: upside risk is a broad August improvement in manufacturing and trade matching July's services rebound, which would land above the interval if the first print is -8.6 or higher. Downside risk is renewed trade/export weakness or cost-pressure pessimism spreading across sectors, which would land below the interval if the first print is -15.2 or lower. Either outside the interval would be a sharper one-month move than the widened recent-change reference class expects.

Review disposition: accepted the required traceability fix by aligning the model-candidate sample, monthly-change list, calibration counts, sigma, and interval explanation to the explicitly fetched 2025-08 through 2026-07 NBB overall-series data. Accepted the resolver-clarity suggestion by stating that the registered NBB business survey page remains the resolving URL while SDMX was used for evidence.

calibrated forecast · 80% CI
-11.9[-15.1 · -8.7]
Target metadata

Data point: nbb.business_barometer.overall.2026-08.first_print

More government data forecasts