Endogeneity in Unconditional Quantile Regression

[Biswa bhusan]

Dear All, I am doing quantile Oaxaca decomposition using unconditional quantile regression proposed by Firpo, S., Fortin, N. M., & Lemieux, T. (2009). Unconditional quantile regressions. Econometrica, 77(3), 953-973. My dependent variable is food security index and one of my independent variable is monthly per-capita expenditure (MPCE) of household. From literature, I found MPCE is endogeneous. Therefore, I followed two step method proposed by Lee, S. (2007). Endogeneity in quantile regression models: A control function approach. Journal of Econometrics, 141(2), 1131-1158. In first step, I estimated control function. In second step, I used control function as an additional variable in the Quantile Oaxaca decomposition. But my question is: Lee method is based on conditional quantile regression, but Firpo method is unconditional quantile regression. So, If I use control function in my decomposition equation, is it possible? Is there any Stata code to address endogeneity for unconditional quantile regression?

[Xinghui Qi]

Dear Biswa bhusan，
when I run an unconditional quantile regression (Firpo, S., Fortin, N. M., & Lemieux, T. 2009) with an endogenous regressor, I have the same problem as you. I'll appreciate it very much if you can give me some help and suggestion. Thank you!

[Shrek Deng]

Hello, have anyone solved this problem? I'm also faced with the endogeneity issue in unconditional quantile regression. Could anyone share the method to deal with this? Thank you.

[FernandoRios]

Hi there,
Unfortunately, there is no one-step solution for problems of endogeneity with UQR. If you search on google about this question, you will most likely be directed to "https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2765035"
This paper suggests that one way to deal with endogeneity is using a control function approach. So perhaps you can use that approach for your research
HTH

[Shrek Deng]

Hi there,
Unfortunately, there is no one-step solution for problems of endogeneity with UQR. If you search on google about this question, you will most likely be directed to "https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2765035"
This paper suggests that one way to deal with endogeneity is using a control function approach. So perhaps you can use that approach for your research
HTH

Thank you for your suggestion.

[Biswa bhusan]

Hi there,
Unfortunately, there is no one-step solution for problems of endogeneity with UQR. If you search on google about this question, you will most likely be directed to "https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2765035"
This paper suggests that one way to deal with endogeneity is using a control function approach. So perhaps you can use that approach for your research
HTH

Thank you for your suggestion. However, is there Stata code available for control variable approach by "https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2765035" or "Rothe, C. (2010). Identification of unconditional partial effects in nonseparable models,Economics Letters109: 171–174."

Particularly, how to generate the control variable suggested by above two papers? If any one can suggest, it will be really helpful.

Thank you