Hi
I have a question regarding the necessity of instrumental variables in a bivariate probit model. From my understanding, in nonlinear settings like the bivariate probit model, identification might be possible even without using an instrumental variable (Greene, 2018), relying instead on the model's distributional assumptions. Is it advisable to run a bivariate probit model without an instrument, or are there specific scenarios where an instrument is essential for the validity of the model's results?
I have a question regarding the necessity of instrumental variables in a bivariate probit model. From my understanding, in nonlinear settings like the bivariate probit model, identification might be possible even without using an instrumental variable (Greene, 2018), relying instead on the model's distributional assumptions. Is it advisable to run a bivariate probit model without an instrument, or are there specific scenarios where an instrument is essential for the validity of the model's results?
