lukestein’s Twitter Archive—№ 13,935

@economeager — “No no, we have INDIVIDUAL data on the key predictor.” – “The highly endogenous one?” — “Yeah, but we instrument for it with [thing that varies at the aggregate level]”