Thanks a lot for your great paper and your exciting package.
For analyses in which the instrument, treatment, or outcome variables are spatially correlated, applying a spatial instrumental variable regression (S-2SLS) can alleviate spatial endogeneity and mitigate potential bias (https://doi.org/10.1017/psrm.2018.61).
Therefore, I would like to ask whether you plan to make your diagnostic package compatible with spatial IV models, such as those implemented in the speht package?
In my case, I would love to use ivDiag to check the First-Stage properties and to calculate SE for my S-2SLS models calculated with the sphetpackage. However, this seems unfeasable given that the relevant functions in ivDiagneither allow externally fitted models nor a spatial lag matrix as inputs.
Is there a workaround and or do you plan to integrate spatial IV models into ivDiag?
Dear Apoorva,
Thanks a lot for your great paper and your exciting package.
For analyses in which the instrument, treatment, or outcome variables are spatially correlated, applying a spatial instrumental variable regression (S-2SLS) can alleviate spatial endogeneity and mitigate potential bias (https://doi.org/10.1017/psrm.2018.61).
Therefore, I would like to ask whether you plan to make your diagnostic package compatible with spatial IV models, such as those implemented in the
spehtpackage?In my case, I would love to use
ivDiagto check the First-Stage properties and to calculate SE for my S-2SLS models calculated with thesphetpackage. However, this seems unfeasable given that the relevant functions inivDiagneither allow externally fitted models nor a spatial lag matrix as inputs.Is there a workaround and or do you plan to integrate spatial IV models into
ivDiag?Thanks a lot for your work again!
best Sebastian