1[2 {"id":"C1","text":"In real time, the AR(1) achieved lower RMSE than the naive forecast at h=1 (1.485 vs 2.010 pp; relative improvement 0.261).","results_keys":["rmse_model_h1","rmse_naive_h1","rel_improvement_rmse_h1"],"status":"supported","caveat":"8 errors only; DM-HLN p=0.18, not significant."},3 {"id":"C2","text":"In real time, the AR(1) achieved lower RMSE than the naive forecast at h=2 (1.400 vs 2.265 pp; relative improvement 0.382).","results_keys":["rmse_model_h2","rmse_naive_h2","rel_improvement_rmse_h2"],"status":"supported","caveat":"8 overlapping errors; DM-HLN p=0.11, not significant."},4 {"id":"C3","text":"The AR(1) also had lower MAE than the naive forecast at both horizons.","results_keys":["mae_model_h1","mae_naive_h1","mae_model_h2","mae_naive_h2"],"status":"supported","caveat":"Small sample, single 2-year window."},5 {"id":"C4","text":"The advantage of the AR(1) over the naive forecast is not statistically significant at the 5% level at either horizon.","results_keys":["supp_dm_hln_pvalue_two_sided_h1","supp_dm_hln_pvalue_two_sided_h2","supp_dm_hln_stat_h1","supp_dm_hln_stat_h2"],"status":"supported","caveat":"DM test has low power and weak justification with T=8."},6 {"id":"C5","text":"Mean absolute revision between first release and the current vintage for the h=1 targets is 0.5625 pp, about a third of the model's h=1 RMSE.","results_keys":["mean_abs_revision_h1_targets","rmse_model_h1"],"status":"supported","caveat":"Recent quarters may still be revised further."},7 {"id":"C6","text":"The AR(1) had smaller squared error than naive at 5 of 8 origins at h=1 and 8 of 8 at h=2.","results_keys":["supp_n_model_wins_sqloss_h1","supp_n_model_wins_sqloss_h2"],"status":"supported","caveat":"h=2 errors overlap across adjacent origins."}8]9