What is the problem:
I am using the post-estimation 'predict' command to estimate the hazard rates for individuals after running a cure model (cureregr) and I get the following error message:

Maximum number of iterations exceeded.
r(498);

What am I doing:
I am conducting an individual-level fertility analysis, using longitudinal event history data, in which I am modelling parity progression (the relative risk of having another child) using a split-population/cure model (cureregr). _t is the duration of the interbirth interval for all second and higher-order births combined. The model contains several control variables, age, region, ses for example and is stratified across mother's birth cohorts.

What does my syntax look like:

xi: cureregr i.age i.region i.ses if cohort ==1, dist(lognormal) class(m) link(logi) sc(i.age i.region i.ses) nolog

predict, haz gen(my_hazard)

What else:
The error only seems to arise for some mother's birth cohorts and not others.