1# Verification checks (P4)2| check | outcome |3|---|---|4| Clean re-run (deleted outputs, re-ran analysis.py): results.json byte-identical, analytic.csv & exclusions.csv md5 identical | PASS |5| analytic.csv rows = n_analysed (1800); distinct firms = n_firms (300) | PASS |6| analytic.csv columns contain no rating_2025 / post-t field | PASS |7| Ledger: n_before − n_removed = n_after each row; chained; starts at 2140, ends at 1800 | PASS |8| se_leverage_cluster > 0; 95% CI contains beta | PASS |9| analytic years 2016–2021 (no 2022 origin with unobserved outcome); no duplicate keys; outcome ∈ {0,1} | PASS |10| default_rate equals mean of analytic default_next; 0 ≤ R² ≤ 1 | PASS |11| Magnitude vs baseline: bivariate leverage slope 0.267 vs adjusted 0.264 — consistent | PASS |12