# Data audit (P1)
- data/canada_gdp_employment.csv SHA-256 f4c8e4976fb5f5f0c326e40bcb4f85b8b22963e5be89e2729e1a430c198f2bed (matches README). data/README.md SHA-256 2000e2e6...fa2f.
- 345 rows x 5 cols (month, gdp_level, gdp_yoy, emp_level, emp_yoy); months 1998-01-01..2026-09-01, consecutive, no duplicates.
- Units: gdp_level chained 2017 $ millions SAAR; emp_level thousands SA; *_yoy percent (4 dp).
- Missing: gdp_level/gdp_yoy missing only 2026-08, 2026-09 (unpublished). emp complete.
- Recomputed yoy from levels (1999+): max abs diff 5.0e-5 for both (rounding). 1998 yoy values are present although README says first 12 months empty; cannot be recomputed from file -> used as given, flagged in data_issues.
- Anomalies: pandemic swings 2020-03..2021-12 (gdp_yoy min -16.21 in 2020-04, max 18.54 in 2021-04; emp_yoy -15.58/+16.09). Handled by excluding those months from model estimation (decided a priori; window precedes first origin's information set 2022-03).
- Availability: the file contains data after each origin's cut-off (it is the current vintage). Handling: the script slices by date per origin (gdp <= t-2, emp <= t-1) and asserts it; ledger columns last_*_date_used document it.