You don't think that the children of religious parents are more likely to be religious than children of atheists?
I would expect that in most secular cultures you'll see something like, say, 60/100 children of religious parents becoming atheists, and 90/100 for children of atheist parents.
These kinds of dynamics would still mean that overall, even if religious people are having more children than atheists, the population is still growing more atheist with each generation.