1. Mini-thread on excess deaths and an error in @OurWorldInData re corona. They added South Korea, which looks like this: 
2. In other words, per the tracker, South Korea has 20% excess deaths for people aged 85+, and has since before the crisis started.
3. @OurWorldInData also thinks Korea had more excess deaths than Germany interwave. (Ignore the last few weeks, the data is laggy.)
4. What's going on here is that the Korean population is rapidly aging as we speak, whereas Germany (and Japan) aged a long time ago.
5. The excess deaths are compared with the 2015-9 average. But in 2015-8, Korea's crude death rate rose 5.4 -> 5.8: data.worldbank.org/indicator/SP.D…
6. Germany is only up 11.3 -> 11.5 in that period - its demographic bulge was in the 19th century, Korea's was in the late 20th.
7. Over this 3-year period, Korea's life expectancy rose 0.6 years, Germany's only 0.25. data.worldbank.org/indicator/SP.D…
8. So it's not that there was some public health crisis in Korea before corona, just the normal demographic transition.
9. In Japan, likewise, crude death rates are rising (10.3 -> 11 in 2015-8), so a 2015-9 baseline would create fictitious excess deaths.
10. In a rapidly aging country like Korea, it's most appropriate to compare to a 2019 baseline, or even try adjusting for early 2020.
11. This is why the first graphic in tweet #1 looks the way it is: aging population -> average age in the 85+ group is up -> deaths are up.
12. This error is not significant in Western countries - Canada's crude death rate in 2015-8 was up 4%, the others up 1-2% or even down.
13. Belgium had a big 1-time rise in 2018 per WB data but it must be an error - no growth in seen in national data. statbel.fgov.be/en/open-data/n…
14. Upshot: @OurWorldInData is doing excellent work on excess deaths in the West - but in Korea the same method produces overestimates. /end
• • •
Missing some Tweet in this thread? You can try to
force a refresh
