# P4 checks

• Clean re-run of out/analysis.py: results.json byte-identical (cmp) — PASS
• analytic rows == n_analysed: PASS
• analytic firms == n_firms: PASS
• no rating_2025 in analytic: PASS
• ledger n_before-n_removed==n_after: PASS
• ledger chained: PASS
• ledger start=n_raw, end=n_analysed: PASS
• se>0: PASS
• CI contains beta: PASS
• 0<=r2<=1: PASS
• default_rate == analytic mean: PASS
• max year in analytic <=2021 (outcome observed by end 2022): PASS
• no duplicate firm-year keys: PASS
• Magnitude vs baseline: default_rate 0.121; beta_leverage 0.264 per unit leverage => ~0.026 per 0.1 leverage, plausible relative to baseline — PASS