Correlated Random effects model to*solve endogeneity problem

[Anirban Adhikary]

Hi,
I have done random effects model to check the impact of time-independent variables. Now I could not find suitable IV to control for endogeneity. After performing correlated random effects proposed by mundlak my results are still holding (I have incorporated mean of time-dependent variables as independent variables). Can I say my results are consistent and unbiased in spite of endogeneity? (ref: Bell, A., & Jones, K. (2015). Explaining fixed effects: Random effects modeling of time-series cross-sectional and panel data. Political Science Research and Methods, 3(01), 133-153.)

[Phil Bromiley]

First, you're more likely to get a helpful answer if you follow the guidelines on asking questions in the FAQ - provide Stata code (using code delimiters), Stata output, and sample data (using dataex).
You have two very different issues - endogeneity and fixed vs random effects. As I understand it, Mundlak does not fix endogeneity (unless it somehow was generated by not specifying the panel control correctly). So, no, I don't think your results are consistent if you have an endogenous variable even if you use Mundlak's approach.

[Sasha Gulabivala]

Hi,

I am also running a correlated random effects (CRE) probit model:

xtprobit saving1 x1 x1bar x2 x2bar i.year, re vce(robust)

I wondered whether there is a test or "fix" for endogeneity, and a method for IV regression (equivalent to ivreg2) that is compatible with xtprobit, re?

Similar query posted here:
http://www.statalist.org/forums/foru...ogeneity-tests

Thanks