[Gretl-users] Re: Running a 2SLS model with an interaction term
Am 18.04.2023 um 20:44 schrieb elfsog(a)yahoo.gr:
What if the product Z1*X2 is correlated with the exogenous variable
X2?
This is moving away from a question about gretl, but likely they will be
correlated indeed. So what? Valid instruments can be correlated among
each other (short from being perfectly collinear or close to that).
cheers
sven
Attachments:
- attachment.htm (text/html — 806 bytes)