# Contract — F6-conf-03

Question: How strongly does the quarterly change in the US unemployment rate respond to quarterly real GDP growth (Okun's law, difference form), 1960Q1–2019Q4, and is the response stable across 1960Q1–1983Q4 vs 1984Q1–2019Q4?

Estimand: β in Δu_t = α + β g_t + ε_t, Δu_t = change in quarterly-average UNRATE (pp), g_t = 100·Δln GDPC1 (% q/q, not annualised). OLS β is a meaningful linear-projection coefficient under stationarity of (Δu, g); it is a reduced-form association, not a structural causal effect. Inference: Newey–West HAC (Bartlett), lag reported; 95% CI from HAC SE (normal critical value). Stability: subperiod OLS slopes + classical Chow SSR F-test of (α,β) equality, F(2, n−4).

Outputs: out/results.json (keys: sample_start, sample_end, n_sub1, n_sub2, baseline_metric, model_metric, effect_estimate, intercept, effect_se, hac_lags, effect_ci_low, effect_ci_high, beta_sub1, beta_sub2, chow_f_stat, chow_p_value, stability_conclusion, conclusion, task_id, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed); out/figure1.png; out/table1.csv (period,n,intercept,slope,slope_se; 3 rows); out/report.md (6 sections); out/analysis.py; out/claims.json.

Information rules: u_t = mean of 3 monthly values; sample = 240 quarters 1960Q1–2019Q4 (1959Q4 levels used only for the first difference); nothing after 2019Q4 in any estimate; only data/GDPC1.csv, data/UNRATE.csv, data/README.md (gate_spec.json/bundle.json are not inputs and are not read, apart from bundle.json hashes not being used either).

Invalidity/impossibility: missing monthly or quarterly values inside 1959Q4–2019Q4; sample count ≠ 240; post-2019 data entering estimates; non-HAC SEs for full-sample β.