RDP 2025-09: Forecasts of Period-average Exchange Rates: Insights from Real-time Daily Data Appendix E: Robustness – Model-based Forecast Assumptions
December 2025
- Download the Paper 942KB
This appendix presents forecasts that use pre-sample testing for order of integration and the number of autoregressive terms using AIC. These tests were applied to the daily ARIMA specification, and the equivalent lag structure was imposed on Equations (5) to (10).
| Forecast | Model inputs | Horizon (months) | |||||
|---|---|---|---|---|---|---|---|
| 1 | 3 | 6 | 12 | 24 | 36 | ||
| RMSFE ratio | |||||||
| Recursive | Month-average | 1.01 | 1.01 | 1.02 | 1.03 | 1.06 | 1.08 |
| Recursive | End-of-month | 0.89 | 0.97 | 0.99 | 1.02 | 1.04 | 1.07 |
| Recursive | Bottom-up | 0.88 | 0.96 | 0.99 | 1.01 | 1.01 | 1.03 |
| Direct | Month-average | 1.02 | 1.01 | 1.01 | 1.03 | 1.07 | 1.24 |
| Direct | End-of-month | 0.88 | 0.97 | 0.99 | 1.01 | 1.07 | 1.24 |
| Direct | UMIDAS | 0.88 | 0.96 | 0.99 | 1.01 | 1.07 | 1.24 |
| Success ratio | |||||||
| Recursive | Month-average | 0.56 | 0.52 | 0.51 | 0.49 | 0.54 | 0.52 |
| Recursive | End-of-month | 0.68 | 0.59 | 0.54 | 0.52 | 0.57 | 0.54 |
| Recursive | Bottom-up | 0.68 | 0.57 | 0.54 | 0.52 | 0.59 | 0.57 |
| Direct | Month-average | 0.56 | 0.51 | 0.51 | 0.50 | 0.55 | 0.54 |
| Direct | End-of-month | 0.68 | 0.57 | 0.53 | 0.52 | 0.56 | 0.54 |
| Direct | UMIDAS | 0.68 | 0.57 | 0.54 | 0.52 | 0.56 | 0.54 |
| Notes: Reports the median result across countries relative to the month-average no-change forecast. RMSFE ratios less than 1 improve upon the month-average no-change. Success ratios greater than 0.5 are improvements upon random chance. | |||||||
| Forecast | Model inputs | versus end-of-month no-change | versus month-average no-change | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| REER | RER | NEER | NER | REER | RER | NEER | NER | |||
| Mean square accuracy (%) | ||||||||||
| Recursive | Month-average | 9 | 3 | 9 | 3 | 18 | 6 | 25 | 26 | |
| Recursive | End-of-month | 45 | 15 | 26 | 1 | 58 | 73 | 54 | 87 | |
| Recursive | Bottom-up | 59 | 32 | 43 | 4 | 68 | 82 | 63 | 96 | |
| Direct | Month-average | 8 | 0 | 8 | 0 | 14 | 4 | 19 | 20 | |
| Direct | End-of-month | 44 | 9 | 24 | 0 | 55 | 73 | 46 | 75 | |
| Direct | UMIDAS | 38 | 8 | 24 | 0 | 51 | 76 | 55 | 84 | |
| Directional accuracy (%) | ||||||||||
| Recursive | Month-average | 14 | 14 | 11 | 14 | 48 | 62 | 51 | 69 | |
| Recursive | End-of-month | 15 | 10 | 19 | 10 | 98 | 89 | 100 | 100 | |
| Recursive | Bottom-up | 28 | 20 | 34 | 32 | 99 | 89 | 100 | 100 | |
| Direct | Month-average | 14 | 13 | 10 | 11 | 41 | 51 | 43 | 58 | |
| Direct | End-of-month | 18 | 6 | 16 | 8 | 98 | 87 | 99 | 95 | |
| Direct | UMIDAS | 38 | 14 | 16 | 12 | 100 | 88 | 99 | 95 | |
| Notes: Reports the per cent of countries for which the model-based forecast one-month-ahead significantly outperforms the benchmark at a five per cent level of significance. | ||||||||||