1[2 {"id":"C1","text":"Over 1960Q1-2019Q4, a 1% higher quarterly real GDP growth rate is associated with a 0.256 pp smaller rise in the unemployment rate; the HAC 95% interval lies entirely below zero.","results_keys":["effect_estimate","effect_se","effect_ci_low","effect_ci_high","hac_lags","conclusion"],"status":"supported","caveat":"Reduced-form association from a static difference model; not causal."},3 {"id":"C2","text":"The Okun regression lowers the in-sample RMSE of the quarterly change in unemployment from 0.325 pp (mean-only baseline) to 0.250 pp.","results_keys":["baseline_metric","model_metric","r_squared_full"],"status":"supported","caveat":"In-sample fit only; no out-of-sample evaluation."},4 {"id":"C3","text":"The classical Chow test rejects joint equality of intercept and slope across 1960Q1-1983Q4 and 1984Q1-2019Q4 at 5% (F = 6.09, p = 0.0026), so the relation is classified as unstable.","results_keys":["chow_f_stat","chow_p_value","stability_conclusion","hac_joint_break_p_value"],"status":"supported","caveat":"Classical Chow assumes i.i.d. errors, which the data contradict; a HAC joint Wald test also rejects."},5 {"id":"C4","text":"The subperiod slopes are nearly identical (-0.259 vs -0.273) and their difference is not statistically significant; the instability is in the intercept/joint relation rather than in the Okun coefficient itself.","results_keys":["beta_sub1","beta_sub2","beta_sub1_hac_se","beta_sub2_hac_se","hac_slope_diff_p_value","intercept_sub1","intercept_sub2","hac_intercept_diff_p_value"],"status":"supported","caveat":"The individual intercept difference is not significant either (HAC p = 0.25); the attribution to the intercept is based on point estimates and the joint test."},6 {"id":"C5","text":"HAC inference on the full-sample beta is robust to the choice of Newey-West lag length between 0 and 12.","results_keys":["hac_se_lag_sensitivity","effect_se"],"status":"supported","caveat":"Only Bartlett kernel with fixed lags examined."}7]8