If you look at the plots of the data sources from the meta-survey of literature that is [0] you'll notice the IFR positively correlated with high population and

If you look at the plots of the data sources from the meta-survey of literature that is [0] you'll notice the IFR positively correlated with high population and infrastructure. These places get a better idea of the actual IFR and are diluted out by places that falsify data (China) and aren't capable of taking real population level data (smaller and undeveloped countries). The first part of the paper is justifying this choice as to not dilute out the measurements from the small regions. I guess if you hang a lampshade on it like that it almost makes sense this paper was approved. But it's easily misinterpreted.