The robust F-statistic as a test for weak instruments
For the linear model with a single endogenous variable, (Montiel Olea and Pflueger 2013) proposed the effective F-statistic as a test for weak instruments in terms of the Nagar bias of the two-stage least squares (2SLS) or limited information maximum likelihood (LIML) estimator relative to a benchma...
| 1. Verfasser: | |
|---|---|
| Format: | Journal article |
| Sprache: | English |
| Veröffentlicht: |
Elsevier
2025
|
| _version_ | 1826317347602825216 |
|---|---|
| author | Windmeijer, F |
| author_facet | Windmeijer, F |
| author_sort | Windmeijer, F |
| collection | OXFORD |
| description | For the linear model with a single endogenous variable, (Montiel Olea and Pflueger 2013) proposed the effective F-statistic as a test for weak instruments in terms of the Nagar bias of the two-stage least squares (2SLS) or limited information maximum likelihood (LIML) estimator relative to a benchmark worst-case bias. We show that their methodology for the 2SLS estimator applies to a class of linear generalized method of moments (GMM) estimators with an associated class of generalized effective F-statistics. The standard robust F-statistic is a member of this class. The associated GMMf estimator, with the extension “f” for first-stage, has the weight matrix based on the first-stage residuals. In the grouped-data IV designs of Andrews (2018) with moderate and high levels of endogeneity and where the robust F-statistic is large but the effective F-statistic is small, the GMMf estimator is shown to behave much better in terms of bias than the 2SLS estimator. |
| first_indexed | 2025-02-19T04:37:03Z |
| format | Journal article |
| id | oxford-uuid:df5106f8-9b1f-4f6c-a890-6b278e223d9d |
| institution | University of Oxford |
| language | English |
| last_indexed | 2025-02-19T04:37:03Z |
| publishDate | 2025 |
| publisher | Elsevier |
| record_format | dspace |
| spelling | oxford-uuid:df5106f8-9b1f-4f6c-a890-6b278e223d9d2025-02-03T09:34:50ZThe robust F-statistic as a test for weak instrumentsJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:df5106f8-9b1f-4f6c-a890-6b278e223d9dEnglishSymplectic ElementsElsevier2025Windmeijer, FFor the linear model with a single endogenous variable, (Montiel Olea and Pflueger 2013) proposed the effective F-statistic as a test for weak instruments in terms of the Nagar bias of the two-stage least squares (2SLS) or limited information maximum likelihood (LIML) estimator relative to a benchmark worst-case bias. We show that their methodology for the 2SLS estimator applies to a class of linear generalized method of moments (GMM) estimators with an associated class of generalized effective F-statistics. The standard robust F-statistic is a member of this class. The associated GMMf estimator, with the extension “f” for first-stage, has the weight matrix based on the first-stage residuals. In the grouped-data IV designs of Andrews (2018) with moderate and high levels of endogeneity and where the robust F-statistic is large but the effective F-statistic is small, the GMMf estimator is shown to behave much better in terms of bias than the 2SLS estimator. |
| spellingShingle | Windmeijer, F The robust F-statistic as a test for weak instruments |
| title | The robust F-statistic as a test for weak instruments |
| title_full | The robust F-statistic as a test for weak instruments |
| title_fullStr | The robust F-statistic as a test for weak instruments |
| title_full_unstemmed | The robust F-statistic as a test for weak instruments |
| title_short | The robust F-statistic as a test for weak instruments |
| title_sort | robust f statistic as a test for weak instruments |
| work_keys_str_mv | AT windmeijerf therobustfstatisticasatestforweakinstruments AT windmeijerf robustfstatisticasatestforweakinstruments |