Endogeneity correction via Gaussian Copulas - generating normal CDF of a non-normal empirical distribution function

[Alexander Schult]

Dear Stata experts,

I am a passionate passive follower in the forum, you guys helped me a lot in the past 2 years of my research!

I am facing now an "apparently" easy task that actually takes, according to the professor, only 3 lines to compute but I am struggling for 3 hrs how..

Its a class about Endogeneity correction, this specific assignment is via Gaussian Copulas approach.

For this, we need to:
1) Compute a dataset with an endogenous regressor correlated with the error term that is non-normally distributed
2) Then we need to generate the inverse of the normal cumulative distribution function (CDF) of this non-normally distributed error term
3) We integrate this into the actual estimation

I am heavily struggling to compute a non-normally distributed variable and then compute from this the inverse of the normal CDF.

I had a look into this post but I am absolutely not sure how to interpret the first two results and whether they give me what I need:
https://www.statalist.org/forums/for...aussian-copula

Any idea dear Stata experts? That would be amazing.


Best regards,
Alexander