You have to understand limitation of this study. This is not a best estimate for CFR or IFR. This is an estimate based on a review of seroprevalence studies. So

You have to understand limitation of this study. This is not a best estimate for CFR or IFR. This is an estimate based on a review of seroprevalence studies. Some of them are complete crap (IFR 0.0%!), a lot of them are biased towards taking samples from people who went through an influenza like illness recently. And even if done perfectly about 5% of people had some antibodies against coronaviruses even way before the pandemic so there is some background level that should be subtracted from the seroprevalence but almost never is.