You're not taking into account that those born in 1920-24 would have had higher death rates their whole lives compared to those born 1945-49. Also you don't take into account that many more migrants to France were born 1945-49 compared to 1920-24.
France population spreadsheet.The spreadsheet covers 1990 to 2020, with the population by one year age groups, except for 5-year groups for 85-89; 90-94; and 95-99; and unlimited for 100+.
The groups born in WWI and WWII are indicated in red. Those born between 1890-1900 who would have born the brunt of WWI are indicated in grey.
For each year, the population is in the first column, the second column is the difference between persons who are age N in year Y, and who were N-1 in Y-1. The difference represents net migration minus deaths.
Under 40 you are seeing in-migration. Past 40 you are mostly seeing deaths, unless there is some net out-migration.
For the 5-year age groups, the difference column compares the group vs. the same group five years earlier when they were 5 years younger, and divides by 25 (5-year group over five years) to calculate an average per year for a one-year group. This makes the number for the 85-89 group somewhat comparable to that for 84 one-year group. It does spread some deaths around but is OK for seeing trends.
In the first column for each year, in Row 92 is the mean age, which increased at relentlessly even pace over the 30 years, as a remarkably steady number of births has occurred since WWII (a period of 45 years by 1990, now extended to 75 years by 1990).
In the second column for each year, there is the estimate number of deaths for those over 40 (actually the net change in population), the mean age of death, number of deaths by five-year groups 50 or over, and deaths from 50-79, and 80+. The important thing to not is that the most deaths occur in the 80s.
In France, the drop in births was much greater during WWI, as potential fathers were being slaughtered. The drop in births was less during WWI since France was largely out of the fighting early, and there had been a decline since the early 30s due to the depression.
The drops during the world wars is easier to see when looking horizontally, as opposed to vertically. When you look vertically, you are comparing persons of different ages in the same year. But larger numbers of older persons have died.
For example, in 1990, there were more people living who were 66 (born in 1924) than were 49 (born in 1941, the WWII minimum). By 2000, when they were 76 and 59, those born in 1941 were 18% more numerous.
In 2005, when they were 64 and 81, those born in 1941 were 38% more numerous, and the older generation was dying off much faster (about 16,000 per year vs. 4,000 per year). By 2008, when they 67 and 84, the younger group was 62% larger. By 2019, the 1941 generation, now 78, will be twice as numerous as the five-year cohort born from 1920 to 1924 (95-99). And the older group will still be dying in higher numbers.
When you look horizontally, you are comparing persons of the same age, in different years. Increases in life expectancy will favor those born at later date, but the increase is measured in fractions of a year of longevity.
For example, in 2000, 1.8% of 70 YO died in the previous year. But 4.9% of those who were 80 died. But if we look at 2010, 1.2% of those who were 70 YO had died.
Both your birth year and age have an effect on your chances of dying, but age is much more significant.
If we look at 76 YO in 1990, who were born in 1914, there 376K. By 1993, this number had dipped to 210K (born in 1917), and by 1996 had rebounded to 367K, and by 1997 (born in 1921) was 452K The number of births in 1917 was about 40% less than the number in 1914 or 1920. Comparing those born in 1917 to those born in 1921 there was a 115% uptick.
If we compare those born in 1941, the low year during WWII, to 1948 the peak year of the baby boom the increase was only 62%. Births had been slowly declining from 1931 onward, likely due to the depression, plus the low numbers from WWI reaching adulthood and parenting age.
WWI was a chasm. WWII was a swale with a small bluff on the post-war side.
From 1995 to 2000 the number of deaths increased slightly from 518K to 527K, and then dropped to 495K by 2007, as the smaller WWI generation passed through their 80's (80-84 in 2000, 87-91 in 2007).
Since 2007 deaths have been increasing and will reach 610K by 2020. The rate of increase was somewhat slow at first, and then reached a peak and has begun to decline. The initial increase was almost entirely due to the post-WWI group reaching their 80s and 90s, and replacing those born in WWI. By 2020, the WWI group will be over 100 and the post-WWI group in their late 90s. With only 200K left of those born from after WWI there won't be much increase from them, so the increase in deaths will slow.
There will be a small increase as the early baby-boomers replace those born in the Depression and WWII, but this won't begin to show up until after 2025.