Pooling of Forecasts.
We consider forecasting using a combination, when no model coincides with a non-constant data generation process (DGP). Practical experience suggests that combining forecasts adds value, and can even dominate the best individual device. We show why this can occur when forecasting models are differen...
| Main Authors: | , |
|---|---|
| Format: | Working paper |
| Language: | English |
| Published: |
Nuffield College (University of Oxford)
2002
|
| _version_ | 1826261207020994560 |
|---|---|
| author | Hendry, D Clements, M |
| author_facet | Hendry, D Clements, M |
| author_sort | Hendry, D |
| collection | OXFORD |
| description | We consider forecasting using a combination, when no model coincides with a non-constant data generation process (DGP). Practical experience suggests that combining forecasts adds value, and can even dominate the best individual device. We show why this can occur when forecasting models are differentially mis-specified, and is likely to occur when the DGP is subject to deterministic shifts. Moreover, averaging may then dominate over estimated weights in the combination. Finally, it cannot be proved that only non-encompassed devices should be retained in the combination. Empirical and Monte Carlo illustrations confirm the analysis. |
| first_indexed | 2024-03-06T19:17:53Z |
| format | Working paper |
| id | oxford-uuid:190a20ea-601d-47a0-9f37-5db1c64106af |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-06T19:17:53Z |
| publishDate | 2002 |
| publisher | Nuffield College (University of Oxford) |
| record_format | dspace |
| spelling | oxford-uuid:190a20ea-601d-47a0-9f37-5db1c64106af2022-03-26T10:46:42ZPooling of Forecasts.Working paperhttp://purl.org/coar/resource_type/c_8042uuid:190a20ea-601d-47a0-9f37-5db1c64106afEnglishDepartment of Economics - ePrintsNuffield College (University of Oxford)2002Hendry, DClements, MWe consider forecasting using a combination, when no model coincides with a non-constant data generation process (DGP). Practical experience suggests that combining forecasts adds value, and can even dominate the best individual device. We show why this can occur when forecasting models are differentially mis-specified, and is likely to occur when the DGP is subject to deterministic shifts. Moreover, averaging may then dominate over estimated weights in the combination. Finally, it cannot be proved that only non-encompassed devices should be retained in the combination. Empirical and Monte Carlo illustrations confirm the analysis. |
| spellingShingle | Hendry, D Clements, M Pooling of Forecasts. |
| title | Pooling of Forecasts. |
| title_full | Pooling of Forecasts. |
| title_fullStr | Pooling of Forecasts. |
| title_full_unstemmed | Pooling of Forecasts. |
| title_short | Pooling of Forecasts. |
| title_sort | pooling of forecasts |
| work_keys_str_mv | AT hendryd poolingofforecasts AT clementsm poolingofforecasts |