Hi all,

I would like to know how to compute LATE after a biprobit model with endogenous binary regressor. The formula is given at the top of p.151 of Angrist and Pischke's Mostly Harmless Econometrics (2008), but I'm not sure on how to compute it with postestimation commands.

I know how to get ATE with Austin Nichols' slides (https://www.stata.com/meeting/chicag...11_nichols.pdf) but the commands to get LATE are not mentioned I believe.

Thanks a lot.