I am receiving a dot for the standard error that relates to a coefficient
in xtlogit
Panels with zero positives do not seem to be the problem as other
estimations with xtlogit that do have some panels with zero positives do
not have a dot in the results.