Another week of French mortality statistics has just been posted on line by France’s stats agency INSEE (ghoulish, moi?) They support my assertion – and I’m pleased to see The Economist/ FT now agree - that Fr. is pretty accurate in its running total of Covid-19 deaths. 1/8
The new figures - insee.fr/fr/information… - cover 1 March/27 April. They give comparisons with 2018 (a bad year for ‘flu) and 2019 (a low ‘flu year). Compared to 2019, the spike of deaths for 2020 is 23,916 – close to official C19 death toll on 27 April which was 23,293. 2/8
Straight comparison with 2019 (a low death year) may be misleading. If you average deaths 2018-19 for the same periods, “excess mortality”, or spike in deaths this year, is 19,897. This is OVER 3,000 FEWER than Fr govt’s estimated toll for C19. Strange? Not necessarily 3/8
The smaller print of the INSEE numbers is worth studying. Compared to the same period in 2019, deaths in France among under 25’s have FALLEN this year by 18% and by 25% for young men. The likely explanation is a big drop in road deaths and other accidents during the lockdown. 4/8
The increased death rate for each tranche of age is also fascinating. People aged 25 to 49: no change on 2019 (which was a low death year remember). Ages 50-64 - 12% more deaths this year. 65-74 - 22% more. 75-84 -31% up, Over 84 – 33% up.
5/8
5/8
There is an anomaly in the INSEE stats, however. Home deaths aren't included in the French C19 figures. The INSEE stats point to a 30% increase in home deaths March 1- April 27, mostly among over 75’s. This is roughly 3,450 extra deaths – missing from the official C19 total 6/8
The anomaly can be explained in various ways, including the reduction in deaths amongst the young. It is also probable that many elderly people who would have died of flu in a “normal” year have succumbed to C19 instead. They therefore don’t show up as “extra deaths”.7/8
CONCLUSION: The “excess mortality” method of computing the “real” C19 death toll is useful but inexact. The INSEE figs suggest that the French total - now standing at almost 26,000 – is more accurate than many other countries, but may still be 3,000-5,000 undercounted. 8/8
