Yes, the birth rate is decreasing, but the survival rate of young children has also never been better. That would mean the number of children each woman is raising to adulthood is changing slower than the birth rate.
It cites Niger as an example where more than 1 in 5 children were dying before the age of 5, and now the number has dropped closer to 1 in 10. So as to avoid dealing with fractional numbers of children, I'm going to use an example of ten mothers, but doesn't this mean that, for a group of ten mothers to have 40 children survive to adulthood, they previously needed to give birth 50 times, but now they only need to give birth 44 times? So the birth rate can fall by more than 10% without affecting the total population in a situation where the childhood survival rate increases?
What would you call a statistic about "children who made it to adulthood, per adult woman"?
It sounds like, at least in some countries, women aren't "choosing to have smaller families" as much as "not having to give birth an additional time to bring the family number back to what it would have been had they not lost a child."