Hi all,
I want to estimate a model where:
- the dependent variable is ordinal (from 0 to 2)
-and I have an endogenous regressor which is a dummy,
to solve the endogeneity problem I decided to use the method of instrumental variables applying the conditional mixed process "cmp by Roodman".
Subsequently, to take into account the fixed effects and the unobserved heterogeneity I would like to estimate the cmp with the random effects method and the Mundlak correction.
Now I have a doubt: should I include the mean of my instrumental variables? If yes, both in the equation of the first stage and the second one?
Advice or study material are welcome.
Thank you.
Best
I want to estimate a model where:
- the dependent variable is ordinal (from 0 to 2)
-and I have an endogenous regressor which is a dummy,
to solve the endogeneity problem I decided to use the method of instrumental variables applying the conditional mixed process "cmp by Roodman".
Subsequently, to take into account the fixed effects and the unobserved heterogeneity I would like to estimate the cmp with the random effects method and the Mundlak correction.
Now I have a doubt: should I include the mean of my instrumental variables? If yes, both in the equation of the first stage and the second one?
Advice or study material are welcome.
Thank you.
Best
Comment