Ah nice. Digging back through the years, the spike doesn't feel as big anymore. For sure, the 2022 increase over 2021 of about 3,600 deaths is an increase not seen ever before in this dataset (which starts after WW2 was already well over), but in recent years the increase has been in the ballpark of 500 extra deaths per year (jumping around a lot in individual years). If we start from say 2018, then we expect about 2,500 more deaths in 2023 vs 2018. 2018 had about 33,000 deaths, so 2023 we would expect about 35,500 compared to the actual 37,800. Hmm, ok 2,000 more than expected.

And if we look at the extras from the year before, but minus the missing people from all the ones that didn't die in 2020 when we were in lockdown, the difference of about 4,000 gets us back in the ballpark of the number of people who have died from COVID.

When I started writing this I thought I was uncovering extra deaths from an unknown cause, but nah seems it's just COVID.

Yeah I don't know. Like I get it at a conceptual level.

It basically says that they work out the fertility rate by dividing the number of women by the number of babies. They say they do it in 5 year blocks, I guess this accounts for say lots of kids or lots of old women skewing the numbers when they don't have babies.

As for why we have 38,000 deaths and 59,000 births and this is below the replacement rate, I start to feel like I understand but then decide I don't. The best answer I've found is a suggestion that births per women below 2.1 doesn't necessarily mean a shrinking population, because of the distribution of women of various ages may mean many women of childbearing age (say, through immigration) can cause the population to grow despite births per women being below 2.1.