1[2 {"id":"C1","text":"Over 48 origins the employment-augmented direct regression has lower RMSE than the naive forecast at h=1 (0.743 vs 0.814; relative improvement 0.087).","results_keys":["rmse_model_h1","rmse_naive_h1","rel_improvement_rmse_h1"],"status":"supported","caveat":"Point estimate only; DM p=0.17; current-vintage data."},3 {"id":"C2","text":"At h=3 the model RMSE is 0.902 vs naive 1.140 (relative improvement 0.208).","results_keys":["rmse_model_h3","rmse_naive_h3","rel_improvement_rmse_h3"],"status":"supported","caveat":"Point estimate only; DM p=0.16."},4 {"id":"C3","text":"Neither improvement over the naive forecast is statistically significant at conventional levels (Diebold-Mariano).","results_keys":["dm_pvalue_two_sided_h1","dm_pvalue_two_sided_h3","dm_stat_h1","dm_stat_h3"],"status":"supported","caveat":"HAC lag choice is a judgement call; 48 overlapping origins."},5 {"id":"C4","text":"The gain over naive is not attributable to employment growth: the same regression without employment has lower RMSE at both horizons, and the employment coefficient is small.","results_keys":["diag_rmse_ar_only_h1","diag_rmse_ar_only_h3","rmse_model_h1","rmse_model_h3","last_coefficients_h1","last_coefficients_h3"],"status":"supported","caveat":"Diagnostic added after inspecting coefficients; no significance test of model vs AR-only."},6 {"id":"C5","text":"All forecasts respect the publication-lag rules (GDP <= t-2, employment <= t-1), confirmed by assertions and an independent masked re-estimation.","results_keys":["last_target_date_available","last_predictor_date_available"],"status":"supported","caveat":"Timing only; not real-time vintage validity."}7]8