Using IV in panel data with binary dependent and independent variables.

[RADHIKA BANSAL]

Hi I am using a panel data and both my dependent and independent variables are binary. I have two questions:

1. Should the probit model be used here?
2. Can I use something like "xtivprobit" ?

[Jeff Wooldridge]

xtivprobit is for a (roughly) continuous endogenous explanatory variable. And it maintains serial independence of the shocks, which is strong. You can use biprobit and cluster your standard errors at the unit level.

[RADHIKA BANSAL]

Thank you so much for your reply. My concern is that my independent variable is binary. Would xtivprobit work for that?