1[2 {"id":"C1","text":"The direct OLS model with employment has lower out-of-sample RMSE and MAE than the naive forecast at h=1 (RMSE 0.743 vs 0.814, ~9% reduction).","results_keys":["rmse_model_h1","rmse_naive_h1","mae_model_h1","mae_naive_h1","rel_improvement_rmse_h1"],"status":"supported","caveat":"Current-vintage data; 48 origins only."},3 {"id":"C2","text":"At h=3 the model reduces RMSE by ~21% relative to naive (0.902 vs 1.140) and also has lower MAE.","results_keys":["rmse_model_h3","rmse_naive_h3","mae_model_h3","mae_naive_h3","rel_improvement_rmse_h3"],"status":"supported","caveat":"Not significant at 5% by Diebold-Mariano."},4 {"id":"C3","text":"The model's advantage over naive is borderline significant at h=1 (DM p=0.048) and not significant at h=3 (p=0.108).","results_keys":["dm_stat_h1","dm_pvalue_two_sided_h1","dm_stat_h3","dm_pvalue_two_sided_h3"],"status":"supported","caveat":"Small-sample DM test with HAC variance; 48 overlapping errors."},5 {"id":"C4","text":"LFS employment growth does not drive the gain: the same regression without employment has slightly lower RMSE at both horizons.","results_keys":["rmse_ablation_no_emp_h1","rmse_ablation_no_emp_h3","rmse_model_h1","rmse_model_h3"],"status":"supported","caveat":"Difference not tested for significance; single specification."},6 {"id":"C5","text":"At the last origin (2026-04-01) the latest usable GDP month is 2026-02 and the latest usable employment month is 2026-03; all forecasts respect these lags.","results_keys":["last_target_date_available","last_predictor_date_available","last_origin"],"status":"supported","caveat":"Timing simulation only, not real-time vintages."}7]8