search

lrberge / fixest

Fixed-effects estimations
https://lrberge.github.io/fixest/
361 stars 59 forks source link

under-identification test for instrumental var #474

Open StarryCatte opened 4 months ago

StarryCatte commented 4 months ago

I'm now using feols() to estimate a regression with instrumental variable. The number of instrumental variable is less than the number of endogenous variable. How can I test whether the equation is under identified, for example, using Kleibergen-Paap rk LM statistics?

kylebutts commented 3 months ago

Are you misspeaking here? A regression with fewer instrumental variables than endogenous variables should not be able to produce a coefficient estimate. The testing procedure is for the over-identified case where the number of instruments is larger than the number of endogenous variables