# Data audit
| file | SHA-256 | rows (excl. header) | cols | range | freq | units |
|---|---|---|---|---|---|---|
| GDPC1.csv | 1c6c6e9f59e383e20c849dd299175c7a307526c09abdc18f18da9754fc5b3343 | 319 | 2 (observation_date, GDPC1) | 1947-01-01 – 2026-07-01 | quarterly | bn chained 2017 $, SAAR |
| UNRATE.csv | 621ff6fadf50910cb7a21ee20ba610c247cb043b29b4e05c4c1a72fadca2a06c | 945 | 2 (observation_date, UNRATE) | 1948-01-01 – 2026-09-01 | monthly | percent, SA |

Both hashes match data/README.md and bundle.json.
- Dates: all parse, all fall on the first day of the period, sorted, no duplicates, no gaps in the date grid.
- GDPC1: no missing values. Range 2172.4–24611.1.
- UNRATE: one missing value (empty), at 2025-10-01. This was presumably not published because of the 2025 federal shutdown. It lies outside the analysis window and is excluded by the 2019Q4 cutoff, so it has no effect on the results. Recorded in data_issues.
- Analysis window: 1959Q4–2019Q4 contains 244 GDP rows from 1959-01, i.e. 241 quarters from 1959Q4. UNRATE has 723 months from 1959-10 to 2019-12. There are no NaNs in the window.
- Availability: both files extend past 2019Q4 (to 2026). The post-2019 rows are dropped before any computation (see exclusions.csv).
- Vintage: this is the 2026 vintage of revised GDP. That is acceptable for an ex post question but is noted as a limitation.