[Gretl-users] on pooled panel estimation
Dear Allin,
While estimating a pooled panel model through the ols command I realized that the F
statistics is different if the --cluster option is used although the reported degrees of
freedom remain unchanged (as it should be).
Besides, does the --cluster option overrides the selection of HAC?
Best,
Artur
Attachments:
- attachment.html (text/html — 922 bytes)