thesisPrototype

Conditional

DoD FY2027 prime award obligations if FY2027 NDAA enacted

What will USAspending API v2 report for Department of Defense agency 097 prime award obligations for fiscal year 2027, in billions USD, using the registered query snapshot field obligations from the FIRST PRINT resolving snapshot?

Conditional on: An Act authorizing appropriations for military activities of the Department of Defense for fiscal year 2027 (the FY2027 NDAA; the House-reported bill is H.R. 8800, 119th Congress, House Report 119-698) is enacted into law on or before 2026-12-31.

Forecast

$545.3B
$495.4B–$595.2B
80% prediction interval
gpt-5.5Run record ↗

Cumulative probability

Chance that the outcome is at or below each value.

0%25%50%75%100%$420.6B$545.3B$670.1BForecast value

Shaded band: 80% interval ($495.4B–$595.2B). Dashed line: point estimate ($545.3B).

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

Analysis

§Registered USAspending FY2027 DoD Prime Award Obligations

The target is the registered USAspending API v2 query for agency 097, field obligations, transformed by 1e-9 into billions USD. The ledger commits resolutionDate 2027-10-22, matching the end of the expected release window 2027-10-15 to 2027-10-22. The slug was checked against app.thesisinstitute.org/specs.json in this run and no duplicate published slug was returned by the local match command.

▸ Reported tool use: shell.curlmodel report
GET registered sourceBinding URL for FY2016-FY2020
↳ USAspending agency 097 awards obligations fetched this run: FY2016 obligations=304733246938.93, FY2017 obligations=327990949917.96, FY2018 obligations=365937501255.11, FY2019 obligations=391333992872.89, FY2020 obligations=431883672152.30; transformed by 1e-9 gives 304.733, 327.991, 365.938, 391.334, 431.884 billions USD.
▸ Reported tool use: shell.curlmodel report
GET registered sourceBinding URL for FY2021-FY2025
↳ USAspending agency 097 awards obligations fetched this run: FY2021 obligations=395332709399.53, FY2022 obligations=424512414112.80, FY2023 obligations=467273435599.16, FY2024 obligations=456945504400.91, FY2025 obligations=501543513238.03; transformed by 1e-9 gives 395.333, 424.512, 467.273, 456.946, 501.544 billions USD.
▸ Reported tool use: shell.curlmodel report
GET registered sourceBinding URL for FY2026 and FY2027 resolver-state check
↳ USAspending agency 097 awards query fetched this run returned FY2026 obligations=298980123138.95, transformed to 298.980 billions USD, but FY2026 is incomplete on 2026-08-13. FY2027 query returned error detail: Field fiscal_year value 2027 is above max 2026.

Base rate / reference class: completed FY2016-FY2025 registered-query obligations have mean 406.7B, range 304.7B to 501.5B, and last completed print 501.5B. Successive completed annual changes are +23.3B, +37.9B, +25.4B, +40.5B, -36.6B, +29.2B, +42.8B, -10.3B, and +44.6B, with mean +21.9B, sigma 27.6B, and range -36.6B to +44.6B. Last-print persistence benchmark for FY2027 is 501.5B; the mean-change trend benchmark is 501.5 + 2*21.9 = 545.3B because FY2026 is not yet a completed print.

Model candidates under thesis_model_candidate_v1, train cutoff FY2025 completed registered query: persistence candidate point=501.5B, p10=451.6B, p50=501.5B, p90=551.4B, 80% interval 451.6B-551.4B, 90% interval 431.0B-572.0B, interval_method=annual-change residual sigma over FY2016-FY2025, calibration_n=9, one-step walk-forward MAE=33.4B. Mean-change trend candidate point=545.3B, p10=495.4B, p50=545.3B, p90=595.2B, 80% interval 495.4B-595.2B, 90% interval 463.1B-627.5B, interval_method=two-year annual-change residual sigma sqrt-horizon, calibration_n=9, one-step walk-forward MAE=24.1B. The trend candidate beats persistence in backtest and is selected.

Prior/update/interval: prior = mean-change trend candidate from completed FY2016-FY2025 registered-query history; historical sample = 10 completed values and 9 annual changes. Adjustment components: +0.0B for current FY2026 partial print because it is not a comparable completed fiscal-year snapshot; +0.0B for the NDAA mechanism because the forecast is conditional on enactment by 2026-12-31 and no separate numeric appropriation-to-prime-award conversion fetched this run clears the update test. Point = FY2025 501.5B + 2*(mean annual change 21.9B) = 545.3B. Interval method = annual-change residual, two-year horizon. sigma = 27.6B from successive completed annual changes; basic half-width = 1.28*sigma = 35.3B; widen by sqrt(2)=1.41 for two fiscal years beyond the last completed print, half-width = 1.28*27.6*1.41 = 49.9B. Implied 80% bounds = 545.3 - 49.9 = 495.4B and 545.3 + 49.9 = 595.2B.

Counter-consideration: downside risk outside the interval would come from a prolonged FY2027 continuing resolution, appropriations timing lag despite NDAA enactment, procurement execution delay, or cancellation-driven award slowdown, which would land below 495.4B. Upside risk outside the interval would come from unusually large shipbuilding, aircraft, munitions, or Ukraine/Indo-Pacific supplemental obligations posting into prime awards before the snapshot, which would land above 595.2B.

Review disposition: accepted the resolver clarification by distinguishing the first successful registered-query snapshot from the fixed 2027-10-22 ledger deadline, and accepted the historical-context completeness fix by carrying the full FY2016-FY2025 registered-query history. Kept the sigma, sqrt(2) horizon widening, and interval bounds unchanged.

Key drivers

  • completed FY2016-FY2025 upward trend
  • FY2025 registered-query high print
  • timely NDAA condition supports normal award cycle
  • FY2026 still incomplete at run date
  • large annual obligation volatility

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
USAspending API v2, agency 097 awards
Resolution date
October 22, 2027· outcome not recorded
Resolution rule
If the condition is satisfied, resolve from the first successful registered-query snapshot captured under the target process for USAspending API v2 agency 097 awards with fiscal_year=2027 at https://api.usaspending.gov/api/v2/agency/097/awards/?fiscal_year=2027, with resolutionDate fixed at the ledger deadline 2027-10-22. Read JSON field obligations, multiply by 1e-9, and resolve as Department of Defense prime award obligations for fiscal year 2027 in billions USD under registered_query_snapshot releasePolicy. Later corrections or changes are irrelevant. Round to the nearest 0.1 billion USD. If the condition is not satisfied, mark the cell unresolved.
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 on the target contract with only minor tightening needed around first-print wording and historical-context completeness.

  • warning resolver: The rule says both "on 2027-10-22 fetch" and "Use the first registered query snapshot," which could read as conflicting if an earlier snapshot appears within the 2027-10-15 to 2027-10-22 window.
  • warning coherence: The reasoning uses FY2016-FY2025 as the historical sample, but the JSON historicalContext field only includes FY2023-FY2025.
  • info optional_suggestion: The Prior/update/interval step is strong; keep the explicit sigma, sqrt(2) horizon widening, and implied bounds.
  • info optional_suggestion: Tail scenarios are concrete enough, but mentioning that NDAA enactment does not guarantee appropriations timing would make the downside tail sharper.

disposition accepted: Review disposition: accepted the resolver clarification by distinguishing the first successful registered-query snapshot from the fixed 2027-10-22 ledger deadline, and accepted the historical-context completeness fix by carrying the full FY2016-FY2025 registered-query history. Kept the sigma, sqrt(2) horizon widening, and interval bounds unchanged.

disposition accepted: Review disposition: accepted the resolver clarification by distinguishing the first successful registered-query snapshot from the fixed 2027-10-22 ledger deadline, and accepted the historical-context completeness fix by carrying the full FY2016-FY2025 registered-query history. Kept the sigma, sqrt(2) horizon widening, and interval bounds unchanged.

disposition not applicable: Review disposition: accepted the resolver clarification by distinguishing the first successful registered-query snapshot from the fixed 2027-10-22 ledger deadline, and accepted the historical-context completeness fix by carrying the full FY2016-FY2025 registered-query history. Kept the sigma, sqrt(2) horizon widening, and interval bounds unchanged.

disposition not applicable: Review disposition: accepted the resolver clarification by distinguishing the first successful registered-query snapshot from the fixed 2027-10-22 ledger deadline, and accepted the historical-context completeness fix by carrying the full FY2016-FY2025 registered-query history. Kept the sigma, sqrt(2) horizon widening, and interval bounds unchanged.

Activity artifacts

Complete original trace

§Registered USAspending FY2027 DoD Prime Award Obligations

The target is the registered USAspending API v2 query for agency 097, field obligations, transformed by 1e-9 into billions USD. The ledger commits resolutionDate 2027-10-22, matching the end of the expected release window 2027-10-15 to 2027-10-22. The slug was checked against app.thesisinstitute.org/specs.json in this run and no duplicate published slug was returned by the local match command.

▸ Reported tool use: shell.curlmodel report
GET registered sourceBinding URL for FY2016-FY2020
↳ USAspending agency 097 awards obligations fetched this run: FY2016 obligations=304733246938.93, FY2017 obligations=327990949917.96, FY2018 obligations=365937501255.11, FY2019 obligations=391333992872.89, FY2020 obligations=431883672152.30; transformed by 1e-9 gives 304.733, 327.991, 365.938, 391.334, 431.884 billions USD.
▸ Reported tool use: shell.curlmodel report
GET registered sourceBinding URL for FY2021-FY2025
↳ USAspending agency 097 awards obligations fetched this run: FY2021 obligations=395332709399.53, FY2022 obligations=424512414112.80, FY2023 obligations=467273435599.16, FY2024 obligations=456945504400.91, FY2025 obligations=501543513238.03; transformed by 1e-9 gives 395.333, 424.512, 467.273, 456.946, 501.544 billions USD.
▸ Reported tool use: shell.curlmodel report
GET registered sourceBinding URL for FY2026 and FY2027 resolver-state check
↳ USAspending agency 097 awards query fetched this run returned FY2026 obligations=298980123138.95, transformed to 298.980 billions USD, but FY2026 is incomplete on 2026-08-13. FY2027 query returned error detail: Field fiscal_year value 2027 is above max 2026.

Base rate / reference class: completed FY2016-FY2025 registered-query obligations have mean 406.7B, range 304.7B to 501.5B, and last completed print 501.5B. Successive completed annual changes are +23.3B, +37.9B, +25.4B, +40.5B, -36.6B, +29.2B, +42.8B, -10.3B, and +44.6B, with mean +21.9B, sigma 27.6B, and range -36.6B to +44.6B. Last-print persistence benchmark for FY2027 is 501.5B; the mean-change trend benchmark is 501.5 + 2*21.9 = 545.3B because FY2026 is not yet a completed print.

Model candidates under thesis_model_candidate_v1, train cutoff FY2025 completed registered query: persistence candidate point=501.5B, p10=451.6B, p50=501.5B, p90=551.4B, 80% interval 451.6B-551.4B, 90% interval 431.0B-572.0B, interval_method=annual-change residual sigma over FY2016-FY2025, calibration_n=9, one-step walk-forward MAE=33.4B. Mean-change trend candidate point=545.3B, p10=495.4B, p50=545.3B, p90=595.2B, 80% interval 495.4B-595.2B, 90% interval 463.1B-627.5B, interval_method=two-year annual-change residual sigma sqrt-horizon, calibration_n=9, one-step walk-forward MAE=24.1B. The trend candidate beats persistence in backtest and is selected.

Prior/update/interval: prior = mean-change trend candidate from completed FY2016-FY2025 registered-query history; historical sample = 10 completed values and 9 annual changes. Adjustment components: +0.0B for current FY2026 partial print because it is not a comparable completed fiscal-year snapshot; +0.0B for the NDAA mechanism because the forecast is conditional on enactment by 2026-12-31 and no separate numeric appropriation-to-prime-award conversion fetched this run clears the update test. Point = FY2025 501.5B + 2*(mean annual change 21.9B) = 545.3B. Interval method = annual-change residual, two-year horizon. sigma = 27.6B from successive completed annual changes; basic half-width = 1.28*sigma = 35.3B; widen by sqrt(2)=1.41 for two fiscal years beyond the last completed print, half-width = 1.28*27.6*1.41 = 49.9B. Implied 80% bounds = 545.3 - 49.9 = 495.4B and 545.3 + 49.9 = 595.2B.

Counter-consideration: downside risk outside the interval would come from a prolonged FY2027 continuing resolution, appropriations timing lag despite NDAA enactment, procurement execution delay, or cancellation-driven award slowdown, which would land below 495.4B. Upside risk outside the interval would come from unusually large shipbuilding, aircraft, munitions, or Ukraine/Indo-Pacific supplemental obligations posting into prime awards before the snapshot, which would land above 595.2B.

Review disposition: accepted the resolver clarification by distinguishing the first successful registered-query snapshot from the fixed 2027-10-22 ledger deadline, and accepted the historical-context completeness fix by carrying the full FY2016-FY2025 registered-query history. Kept the sigma, sqrt(2) horizon widening, and interval bounds unchanged.

calibrated forecast · 80% CI
$545.3B[$495.4B · $595.2B]
Target metadata

Data point: usaspending.dod.prime_award_obligations.2027.registered_query_snapshot.fy27_ndaa_enacted

More conditional forecasts