Hello all,

I am now self-learning how to use the correlated random effects model from the IZA conference note that Dr. Wooldridge presented (http://conference.iza.org/conference...nonlin_iza.pdf). In this note, it is shortly mentioned that "CRE may not be able to estimate the "causal" effect. I would like to know what are the requirements/circumstances that CRE allows to measure the causal effect.

Thanks,