Logit predicts failure perfectly interpretation

[Carlos Ingeberg]

Hello statalist, i have a logit model that predicts failure perfectly for some years (years is my independent variable). How should i interpret this? I figured its because all the observations on my dependent variable for some years only take the value 0, but how do i interpret it substantially? Also, since its a perfect prediction i dont get any p-values. Can i interpret anything with regards to the level of significance?

Kind Regards
Carlos

[Carlo Lazzaro]

Carlos:
no variations there; hence MLE gasps.
Usually Stata rules out the set of cuplrit observations.
You have to live with the rest of them.