Hello,
I am working with a firm-level panel running a dynamic GMM-SYS.
Two of the regressors (x1 and x2) determine a sample selection (I keep only firms where x1 and x2 are positive).
Do you think I can correct this sample bias including inverse Mills ratio from a bivariate probit (on x1 and x2) in the GMM-SYS? In this case, am I going to consider it as exogenous?
Thank you in advance for your help,
Mariacristina
I am working with a firm-level panel running a dynamic GMM-SYS.
Two of the regressors (x1 and x2) determine a sample selection (I keep only firms where x1 and x2 are positive).
Do you think I can correct this sample bias including inverse Mills ratio from a bivariate probit (on x1 and x2) in the GMM-SYS? In this case, am I going to consider it as exogenous?
Thank you in advance for your help,
Mariacristina