I agree with Marie: the presence of externalities is in fact very easy
to understand when we think at the "negative" effects on society when
a child is raised poorly. The Nobel Prize J. Heckman has showed
empirically that the success in life, and also the economic
performance, of a child depends mostly on the quality of family life
in the first 5 years. If in the very early stages of life the
education process does not work, the whole society is negatively
affected by that. I still think that the concept of "merit" good
captures better than public or private good the economic side of
raising (well or bad) a child, because the full cost and benefit of a
"new person" goes surely beyond the private domain of the family. The
concrete reality of every society says that most clearly, when States
supply nursery and primary school at a price lower than its full cost:
in welfare economics this is a consequence of considering something as
a merit good. Anyway, I am convinced that neoclassical economics is
badly equipped for understanding the social value of children.

Luigino Bruni