I described the approach in my first comment: for 2018 (now it's over) look at the death rate and then apply estimate it uniformly going forward for all recorded births during the year until the surviving population goes to nil. The death rate is all the people who died, whether 1 jan or 31 dec regardless of the age of the deceased.

If you want greater resolution on a process like this you'll do better to search yourself than to reply on HN commenters like me.