heckman with multinomial selection

[Ammar Tyabji]

Hi everyone,

I am trying to adopt the Heckman two-step model to correct for sample selection but I do not seem to have the correct codes. the selection decision is multinomial: not working, involuntary unemployment, employment in informal sector and employment in formal sector. I want the outcome variable, the log of working hours for only those in the formal sector.
does anyone know the codes for this or is it not possible?

thank you in advance!

[George Ford]

https://www.stata.com/meeting/mexico10/mex10sug_huesca.pdf