Endogeneity and sample selection problem in Multinomial Logit model?

[ram shrestha]

I am interested in effect of employment status (inactive/unemployed/employed) in migration intention (no intention /within country/abroad). Here, my dependent variable: migration intention has three categories. So, I am thinking of using Multinational Logit model. However, I suspect that there is problem of self-selection (endogeneity). The employment status seems to be endogenous. Further, I am including education level as explanatory variable. Education may affect both employment status and migration intention. Hence, there seems to be problem of both self-selection and edogeniety to me. However, I am not sure
So, My question is:
Is there serious problem of selection and endogeniety? If there is problem, since, My model is based on MNL, how do I solve the problem.

Please help me out !!

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

We can't tell you if there is a selection or endogeneity problem in your data. That is a substantive issue specific to your data. GSEM or cmp are likely to handle your problem.

[ram shrestha]

Thank you for reply.
I suspect the problem of endogeniety. However, the problem I am facing is:
In two-stage IV estimation, both stage 1 and stage II equations are taking multinational form (discrete choice with more than two choices).
Is there any stata command or reference related to it?
Please help