# Contract — F6-conf-02

Question. After the Bank of Canada's first emergency 50 bp cut (announced 2020-03-04), how did the GoC 5-year benchmark yield move?

Estimand. mean(Δy_t, post_cut=1) − mean(Δy_t over the 120 trading days immediately preceding the first post_cut=1 row), Δy_t = 100·(y_t − y_{t−1}) bp between consecutive file rows; 95% Welch (unequal-variance) interval. Meaningful only if (a) the file contains rows with post_cut=1, (b) it contains ≥120 rows (plus one lag) before the first such row, (c) the treatment flag is as given. It is a descriptive pre/post contrast, not a causal effect unless nothing else changed (COVID shock coincides — major confounder).

Outputs. out/results.json (sample_start, sample_end, n_treated, baseline_metric, model_metric, effect_estimate, effect_ci_low, effect_ci_high, conclusion, task_id, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed); out/figure1.png; out/table1.csv (year,n_days,mean_yield,mean_change_bp); out/report.md (6 sections); out/analysis.py; out/claims.json; plus protocol files (state/, exclusions.csv).

Information rules. Changes between consecutive rows only; post_cut used exactly as provided, never redefined; only data/gc5y_daily.csv used; audit before estimating; report problems in data_issues.

Impossibility. If there are no post_cut=1 rows (or no pre-window), the estimand cannot be computed → conclusion premise_invalid, feasibility infeasible, null for inestimable keys; no proxy substitution.