An Open-model Forecast-error Taxonomy.
We develop forecast-error taxonomies when there are unmodeled variables, forecast ‘off-line’. We establish three surprising results. Even when an open system is correctly specified in-sample with zero intercepts, despite known future values of strongly exogenous variables, changes in dynamics can...
| Main Authors: | , |
|---|---|
| פורמט: | Working paper |
| שפה: | English |
| יצא לאור: |
Department of Economics (University of Oxford)
2011
|
| _version_ | 1826282437584355328 |
|---|---|
| author | Hendry, D Mizon, G |
| author_facet | Hendry, D Mizon, G |
| author_sort | Hendry, D |
| collection | OXFORD |
| description | We develop forecast-error taxonomies when there are unmodeled variables, forecast ‘off-line’. We establish three surprising results. Even when an open system is correctly specified in-sample with zero intercepts, despite known future values of strongly exogenous variables, changes in dynamics can induce forecast failure when they have non-zero means. The additional impact on forecast failure of incorrectly omitting such variables depends only on shifts in their means. With no such shifts, there is no reduction in forecast failure from forecasting unmodeled variables relative to omitting them in 1-step or multi-step forecasts. Simulation illustrations confirm these results. |
| first_indexed | 2024-03-07T00:43:51Z |
| format | Working paper |
| id | oxford-uuid:83f5c0b5-f9c8-47b5-a795-d00f500c1c04 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T00:43:51Z |
| publishDate | 2011 |
| publisher | Department of Economics (University of Oxford) |
| record_format | dspace |
| spelling | oxford-uuid:83f5c0b5-f9c8-47b5-a795-d00f500c1c042022-03-26T21:47:47ZAn Open-model Forecast-error Taxonomy.Working paperhttp://purl.org/coar/resource_type/c_8042uuid:83f5c0b5-f9c8-47b5-a795-d00f500c1c04EnglishDepartment of Economics - ePrintsDepartment of Economics (University of Oxford)2011Hendry, DMizon, GWe develop forecast-error taxonomies when there are unmodeled variables, forecast ‘off-line’. We establish three surprising results. Even when an open system is correctly specified in-sample with zero intercepts, despite known future values of strongly exogenous variables, changes in dynamics can induce forecast failure when they have non-zero means. The additional impact on forecast failure of incorrectly omitting such variables depends only on shifts in their means. With no such shifts, there is no reduction in forecast failure from forecasting unmodeled variables relative to omitting them in 1-step or multi-step forecasts. Simulation illustrations confirm these results. |
| spellingShingle | Hendry, D Mizon, G An Open-model Forecast-error Taxonomy. |
| title | An Open-model Forecast-error Taxonomy. |
| title_full | An Open-model Forecast-error Taxonomy. |
| title_fullStr | An Open-model Forecast-error Taxonomy. |
| title_full_unstemmed | An Open-model Forecast-error Taxonomy. |
| title_short | An Open-model Forecast-error Taxonomy. |
| title_sort | open model forecast error taxonomy |
| work_keys_str_mv | AT hendryd anopenmodelforecasterrortaxonomy AT mizong anopenmodelforecasterrortaxonomy AT hendryd openmodelforecasterrortaxonomy AT mizong openmodelforecasterrortaxonomy |