Hi
am working with a model that involves three first-stage equations, where each first-stage equation is a probit model, and the second stage is an OLS regression. I am unsure about the correct approach to estimate this and properly account for endogeneity. this is what i found on internat im not sure if this is correct
am working with a model that involves three first-stage equations, where each first-stage equation is a probit model, and the second stage is an OLS regression. I am unsure about the correct approach to estimate this and properly account for endogeneity. this is what i found on internat im not sure if this is correct
HTML Code:
* First Stage: Estimate three probit models probit E1 Z1 X1 X2 X3 predict double p_hat1 gen residuals1 = E1 - p_hat1 probit E2 Z2 X1 X2 X3 predict double p_hat2 gen residuals2 = E2 - p_hat2 probit E3 Z3 X1 X2 X3 predict double p_hat3 gen residuals3 = E3 - p_hat3 * Second Stage: OLS with control functions reg Y E1 E2 E3 residuals1 residuals2 residuals3 X1 X2 X3, robust
