Imperfect IV

[Manini Ojha]

Hi, I am trying to use the -imperfectiv- command (Nevo & Rosen 2012) for a linear model with endog x and instrument z which I believe is correlated with the error as well. How can I include state fixed effects in my regressors with this command?

[Andrew Musau]

I am not familiar with imperfectiv, which is from SSC. If the command allows regressors, you can include state fixed effects using state indicators, i.e.,

i.state

assuming that the command supports factor variables. If not, use the -xi- prefix:

xi: imperfectiv ... i.state

where "state" is a categorical variable containing the list of states.

[Manini Ojha]

Thank you for your response. The command does not support factor variables. I was able to execute the command using the state indicators manually.