1[2 {"id":"C1","text":"At h=1 the employment-augmented direct model reduces RMSE relative to the naive forecast by about 8.7% (0.743 vs 0.814 pp).","results_keys":["rmse_model_h1","rmse_naive_h1","rel_improvement_rmse_h1"],"status":"supported","caveat":"DM p≈0.048 (borderline); gain shrinks to ~1% without the COVID exclusion from training."},3 {"id":"C2","text":"At h=3 the model reduces RMSE relative to the naive forecast by about 20.8% (0.902 vs 1.140 pp).","results_keys":["rmse_model_h3","rmse_naive_h3","rel_improvement_rmse_h3"],"status":"supported","caveat":"DM p≈0.105, not significant at 5% with 48 overlapping origins."},4 {"id":"C3","text":"The model also has lower MAE than naive at both horizons.","results_keys":["mae_model_h1","mae_naive_h1","mae_model_h3","mae_naive_h3"],"status":"supported","caveat":"Current-vintage data; not real-time accuracy."},5 {"id":"C4","text":"All forecasts respect the publication lags (GDP ≤ t−2, employment ≤ t−1); at the last origin these are 2026-02 and 2026-03.","results_keys":["last_target_date_available","last_predictor_date_available"],"status":"supported","caveat":"Timing simulation only; revisions ignored."}6]7