thesisPrototype

Government data

SSI total recipients, Aug. 2026

Social Security Administration SSI Monthly Statistics Table 2 total number of recipients for August 2026, all federally administered payments, end-of-month not seasonally adjusted administrative count, first print.

Overdue · was due Oct 4, 2026

This target was registered with a generic source link rather than a resolver binding, and a registration cannot be changed. The resolver will not substitute a different route to the number.

REGISTERED_WITHOUT_EXECUTOR · resolver status as of Oct 7, 2026

Forecast

7.3M

Point projection

gpt-5.5Run record ↗

Analysis

§Forecast for SSA SSI total recipients, August 2026

Framing and exact resolver: this targets SSA SSI Monthly Statistics Table 2, All Federally Administered Payments, the August 2026 row and Total number of recipients column, reported as an end-of-month administrative count and resolved on the first print. The target registry binds resolutionDate 2026-10-04 and resolutionSourceUrl to the July 2026 Table 2 URL; I preserve that bound URL while forecasting the registered August 2026 dataPointId.

▸ Reported tool use: local.registrymodel report
Inspected records/targets/2026-08-12-9e48da0da68cd2dfb368416d94e9714b1c294f8fcbee6ed27a1720e493ef0300.json and site/src/data/ledger-targets.generated.ts; attempted https://app.thesisinstitute.org/specs.json.
↳ Fetched local public target fields: slug ssi-recipients-august-2026, unit millions, dataPointId ssa.ssi.total_recipients.2026-08.first_print, registered resolutionDate 2026-10-04, expectedReleaseWindow 2026-09-26 to 2026-10-04, sourceBinding transform operation multiply factor 0.001, and sourceBinding URL https://www.ssa.gov/policy/docs/statcomps/ssi_monthly/2026-07/table02.html; app specs fetch returned HTTP 404 during this run.
▸ Reported tool use: ssa.pagemodel report
Opened SSA SSI Monthly Statistics index and publishing schedule pages.
↳ Fetched official timing evidence: the current SSA SSI Monthly Statistics index is July 2026 and says released August 2026; it lists Table 2 as Number of recipients by type of payment, total payments, and average monthly payment; SSA Publishing Schedule lists SSI Monthly Statistics with frequency Monthly; the target bound remains 2026-10-04 because the registry provides a latest expected by-date rather than an exact official day.
▸ Reported tool use: ssa.table02model report
Opened SSA SSI Monthly Statistics July 2026 Table 2 and read the Total number of recipients values.
↳ Fetched Table 2 total recipients after the registered 0.001 transform into millions: 2025-07 7.394877; 2025-08 7.405735; 2025-09 7.436689; 2025-10 7.393406; 2025-11 7.391176; 2025-12 7.392201; 2026-01 7.369510; 2026-02 7.358913; 2026-03 7.351825; 2026-04 7.338132; 2026-05 7.322937; 2026-06 7.323731; 2026-07 7.300297.
▸ Reported tool use: model.candidatesmodel report
Computed thesis_model_candidate_v1 candidates from the fetched July 2025-July 2026 SSA Table 2 post-transform levels in millions.
↳ thesis_model_candidate_v1 persistence: point 7.300297, p10 7.276333, p50 7.300297, p90 7.324261, 80% interval 7.276333-7.324261, 90% interval 7.269505-7.331089, interval_method residual_changes_sigma, calibration_n 12, train_cutoff 2026-07, walk_forward_mae 0.014003, walk_forward_rmse 0.018775. Expanding-drift candidate: point 7.292415, p10 7.268451, p50 7.292415, p90 7.316379, 80% interval 7.268451-7.316379, 90% interval 7.261622-7.323208, interval_method residual_changes_sigma, calibration_n 12, train_cutoff 2026-07, walk_forward_mae 0.014051, walk_forward_rmse 0.022720.

Base rate/reference class: the trailing 13 same-table first-print post-transform levels from July 2025 through July 2026 range from 7.300297 million to 7.436689 million, with mean level 7.367648 million. The latest-value persistence prior is 7.300297 million; recent successive changes have mean -0.007882 million, median -0.008843 million, and the 2026 sequence has fallen from 7.369510 million in January to 7.300297 million in July.

Benchmark and update: persistence slightly beats the expanding-drift rule in this short walk-forward window, so it is the selected candidate. Current evidence pulls mildly downward because the January-July 2026 same-source decline is 0.069213 million, while the only directly fetched July-to-August seasonal comparison, 2025-07 to 2025-08, was +0.010858 million; those offset enough that the rounded August forecast stays at 7.300 million.

Prior/update/interval: prior is selected latest-value persistence at 7.300297 million using the fetched July 2025-July 2026 Table 2 post-transform sample. Adjustment components are -0.007882 million from the trailing mean monthly change, plus +0.010858 million from the fetched 2025 July-to-August same-month seasonal move, damped 25% because persistence won walk-forward MAE, giving final unrounded point 7.301041 million before display rounding to 7.300. Interval method is residual_changes_sigma on the 12 fetched successive monthly changes: sigma = 0.018722 million, so 80% half-width is 1.28*sigma = 1.28*0.018722 = 0.023964 million; 7.300297 - 0.023964 = 7.276333 and 7.300297 + 0.023964 = 7.324261, rounded to 7.276 to 7.324.

Counter-consideration: upside risk outside the interval would require a rebound above 7.324 million, such as a July drop that fully reverses plus another positive August seasonal intake month like 2025. Downside risk outside the interval would require August to continue the January-July downtrend at more than about two recent-sigma monthly moves, landing below 7.276 million.

Review disposition: accepted the blocking resolver critique by replacing the divide-by-1,000,000 wording with the registered sourceBinding transform, operation multiply factor 0.001, while preserving unit millions, dataPointId, source URL, first-print policy, and registered resolutionDate. Also clarified that historical values and model inputs are post-transform millions and stated the final unrounded point in the interval step.

Key drivers

  • latest official July 2026 total is 7.300 million recipients
  • January-July 2026 sequence declined by about 0.069 million recipients
  • one-step walk-forward error is marginally lower for persistence than expanding drift
  • July-August 2025 seasonal move was positive, partly offsetting recent downward momentum
  • monthly administrative churn dominates short-horizon uncertainty

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
Social Security Administration SSI Monthly Statistics
Resolution date
October 4, 2026· outcome not recorded
Resolution rule
Resolve to the first published August 2026 value in SSA SSI Monthly Statistics Table 2, row August 2026, column Total number of recipients for all federally administered payments. The resolving observation is the first August 2026 Table 2 print whenever SSA posts it; 2026-10-04 is the registered latest expected by-date from the Thesis target expectedReleaseWindow, not a claimed exact release day. The registered bound source URL is the July 2026 Table 2 page that authenticates the table structure; use the analogous first-posted August 2026 Table 2 row when SSA publishes it. For this registered target, take the sourceBinding field ssa.ssi.total_recipients and apply the registered transform exactly: operation multiply, factor 0.001; the resulting target value is in millions. Round forecast display values to three decimals. Use the first official print only and ignore later revisions or adjusted checks after the first August 2026 table is posted.
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 in structure but needs resolver-unit wording corrected to match the registered sourceBinding transform.

  • blocking resolver: The resolutionRule says SSA publishes counts as whole recipients and converts by dividing by 1,000,000, but the ledger sourceBinding specifies transform multiply by 0.001 for the registered field.
  • info optional_suggestion: In the Prior/update/interval step, state the final unrounded point before display rounding so the damped adjustment is auditable.
  • info optional_suggestion: Clarify whether the cited historical values are post-transform millions or raw SSA table values.

disposition accepted: Review disposition: accepted the blocking resolver critique by replacing the divide-by-1,000,000 wording with the registered sourceBinding transform, operation multiply factor 0.001, while preserving unit millions, dataPointId, source URL, first-print policy, and registered resolutionDate. Also clarified that historical values and model inputs are post-transform millions and stated the final unrounded point in the interval step.

disposition not applicable: Review disposition: accepted the blocking resolver critique by replacing the divide-by-1,000,000 wording with the registered sourceBinding transform, operation multiply factor 0.001, while preserving unit millions, dataPointId, source URL, first-print policy, and registered resolutionDate. Also clarified that historical values and model inputs are post-transform millions and stated the final unrounded point in the interval step.

disposition not applicable: Review disposition: accepted the blocking resolver critique by replacing the divide-by-1,000,000 wording with the registered sourceBinding transform, operation multiply factor 0.001, while preserving unit millions, dataPointId, source URL, first-print policy, and registered resolutionDate. Also clarified that historical values and model inputs are post-transform millions and stated the final unrounded point in the interval step.

Activity artifacts

Complete original trace

§Forecast for SSA SSI total recipients, August 2026

Framing and exact resolver: this targets SSA SSI Monthly Statistics Table 2, All Federally Administered Payments, the August 2026 row and Total number of recipients column, reported as an end-of-month administrative count and resolved on the first print. The target registry binds resolutionDate 2026-10-04 and resolutionSourceUrl to the July 2026 Table 2 URL; I preserve that bound URL while forecasting the registered August 2026 dataPointId.

▸ Reported tool use: local.registrymodel report
Inspected records/targets/2026-08-12-9e48da0da68cd2dfb368416d94e9714b1c294f8fcbee6ed27a1720e493ef0300.json and site/src/data/ledger-targets.generated.ts; attempted https://app.thesisinstitute.org/specs.json.
↳ Fetched local public target fields: slug ssi-recipients-august-2026, unit millions, dataPointId ssa.ssi.total_recipients.2026-08.first_print, registered resolutionDate 2026-10-04, expectedReleaseWindow 2026-09-26 to 2026-10-04, sourceBinding transform operation multiply factor 0.001, and sourceBinding URL https://www.ssa.gov/policy/docs/statcomps/ssi_monthly/2026-07/table02.html; app specs fetch returned HTTP 404 during this run.
▸ Reported tool use: ssa.pagemodel report
Opened SSA SSI Monthly Statistics index and publishing schedule pages.
↳ Fetched official timing evidence: the current SSA SSI Monthly Statistics index is July 2026 and says released August 2026; it lists Table 2 as Number of recipients by type of payment, total payments, and average monthly payment; SSA Publishing Schedule lists SSI Monthly Statistics with frequency Monthly; the target bound remains 2026-10-04 because the registry provides a latest expected by-date rather than an exact official day.
▸ Reported tool use: ssa.table02model report
Opened SSA SSI Monthly Statistics July 2026 Table 2 and read the Total number of recipients values.
↳ Fetched Table 2 total recipients after the registered 0.001 transform into millions: 2025-07 7.394877; 2025-08 7.405735; 2025-09 7.436689; 2025-10 7.393406; 2025-11 7.391176; 2025-12 7.392201; 2026-01 7.369510; 2026-02 7.358913; 2026-03 7.351825; 2026-04 7.338132; 2026-05 7.322937; 2026-06 7.323731; 2026-07 7.300297.
▸ Reported tool use: model.candidatesmodel report
Computed thesis_model_candidate_v1 candidates from the fetched July 2025-July 2026 SSA Table 2 post-transform levels in millions.
↳ thesis_model_candidate_v1 persistence: point 7.300297, p10 7.276333, p50 7.300297, p90 7.324261, 80% interval 7.276333-7.324261, 90% interval 7.269505-7.331089, interval_method residual_changes_sigma, calibration_n 12, train_cutoff 2026-07, walk_forward_mae 0.014003, walk_forward_rmse 0.018775. Expanding-drift candidate: point 7.292415, p10 7.268451, p50 7.292415, p90 7.316379, 80% interval 7.268451-7.316379, 90% interval 7.261622-7.323208, interval_method residual_changes_sigma, calibration_n 12, train_cutoff 2026-07, walk_forward_mae 0.014051, walk_forward_rmse 0.022720.

Base rate/reference class: the trailing 13 same-table first-print post-transform levels from July 2025 through July 2026 range from 7.300297 million to 7.436689 million, with mean level 7.367648 million. The latest-value persistence prior is 7.300297 million; recent successive changes have mean -0.007882 million, median -0.008843 million, and the 2026 sequence has fallen from 7.369510 million in January to 7.300297 million in July.

Benchmark and update: persistence slightly beats the expanding-drift rule in this short walk-forward window, so it is the selected candidate. Current evidence pulls mildly downward because the January-July 2026 same-source decline is 0.069213 million, while the only directly fetched July-to-August seasonal comparison, 2025-07 to 2025-08, was +0.010858 million; those offset enough that the rounded August forecast stays at 7.300 million.

Prior/update/interval: prior is selected latest-value persistence at 7.300297 million using the fetched July 2025-July 2026 Table 2 post-transform sample. Adjustment components are -0.007882 million from the trailing mean monthly change, plus +0.010858 million from the fetched 2025 July-to-August same-month seasonal move, damped 25% because persistence won walk-forward MAE, giving final unrounded point 7.301041 million before display rounding to 7.300. Interval method is residual_changes_sigma on the 12 fetched successive monthly changes: sigma = 0.018722 million, so 80% half-width is 1.28*sigma = 1.28*0.018722 = 0.023964 million; 7.300297 - 0.023964 = 7.276333 and 7.300297 + 0.023964 = 7.324261, rounded to 7.276 to 7.324.

Counter-consideration: upside risk outside the interval would require a rebound above 7.324 million, such as a July drop that fully reverses plus another positive August seasonal intake month like 2025. Downside risk outside the interval would require August to continue the January-July downtrend at more than about two recent-sigma monthly moves, landing below 7.276 million.

Review disposition: accepted the blocking resolver critique by replacing the divide-by-1,000,000 wording with the registered sourceBinding transform, operation multiply factor 0.001, while preserving unit millions, dataPointId, source URL, first-print policy, and registered resolutionDate. Also clarified that historical values and model inputs are post-transform millions and stated the final unrounded point in the interval step.

calibrated forecast · 80% CI
7.3M[7.3M · 7.3M]
Target metadata

Data point: ssa.ssi.total_recipients.2026-08.first_print

More government data forecasts