I have a problem with instrumental variable estimation. I have a model
where my dependent variable (y) is binary (0 or 1), but my right hand
side endogenous variable is lower censored to 0. I checked the stata
web site and find a program called "divprob" that would work when the
dependent variable is binary, however, this program will yield
inconsistent results if the endogenous variable is not continuous. Is
there any other commands in Stata that will gives me consistent results?
Does anyone has any suggestions regarding this type of model?
