I have read that when one adopts Random Effects (RE) there is no need to include cluster-robust standard errors because of the structure and functioning of the RE model itself.
Could anyone please elucidate on this - or indicate good references to better understand the rationale?
Thank you so much.