Hi everyone,
I have a balanced panel data N=14 (country) T=5. I have developed a Correlated random-effects approach because the random effects assumptions are violated and I would like to investigate the effects of the time-constant variables on the response one. But, I think that some covariates are likely to correlate with the response variable, and there is a problem of endogeneity.
How can I solve this issue?