double-hurdle model not feasible

[Sa Fe]

Hi,

I'm using a panel double-hurdle model using

xtdhreg

command. But I get the following error:

Obtaining starting values for full model:

Iteration 0:   log likelihood =  230358.78
Iteration 1:   log likelihood =  230440.79
Iteration 2:   log likelihood =  230440.81

Fitting full model:

initial values not feasible
r(1400);

Your help is appreciated. Thanks,

[William Lisowski]

Since you don't show us the command you used, nor tell us anything about your data, the advice we can offer is limited to the usual: when a complicated model fails, start with a much simpler model (fewer variables, simpler options, ...) and if that is successful, build your way up to your target model a step at a time, and perhaps you can figure out a variable (or a few variables) that are causing problems.

I assume you have reviewed the original Stata Journal article describing the software and its use.

https://journals.sagepub.com/doi/pdf...867X1401400405