One popular story about the decline in family size over the last two centuries goes like this: Back in the old days, having kids paid.  Children started working when they were quite young, and provided for their parents in their old age.  Then industrialization and/or the welfare state came along and changed everything.  Young children ceased to contribute much economically to their families, and once Social Security, Medicare, and so on were in place, people stopped supporting their aging parents.

It turns out that this story is only half true.  Yes, in the modern era, people give little financial assistance to their elders; even in late adulthood, old-to-young transfers remain larger than young-to-old transfers.  The flaw in the story is the assumption that things used to be different.  In an eye-opening 1996 JEL piece, Ted Bergstrom summarizes evidence showing that even in pre-modern societies, kids did not pay.

Kids did not pay in hunter-gatherer societies:

[A]mong hunter- gatherers, resources flow from older to younger generations and not the other way around. These tribes all had very high average fertility (about eight births per woman), but in each case, children consumed more food than they caught, at all ages from birth until age 18.  Grandparents continued to work hard to support their grandchildren and produced more than they ate. At almost no time in their adult lives, did adults produce less than they consumed. When people became too old and frail to work, death followed quickly. Suicide and euthanasia of the enfeebled were frequently reported.

Kids did not pay in agricultural societies:

Calculations by Mueller and by Goran Ohlin (1969) indicate that a parent who gave birth at age 20 and supported a child from age one to age 15 would receive a monetary rate of return of less than one percent on her investment if she retired at age 60 and was supported by the child until age 85 at the level of living that is normal for old people in peasant societies. When one accounts for the probability that either parent or child may die before the parent reaches 85 years of age, the expected rate of return becomes negative. In a peasant society, where land ownership is possible and where there are markets for borrowing and lending, such low rates of return are not likely to be acceptable on purely financial grounds.

An anti-natalist might take this as further proof that breeding is sheer idiocy.  To me, though, it confirms that an intrinsic or “consumption” demand for kids is deeply rooted in human nature.  It also shows that explaining the long-run decline in family size is harder than it looks.  If parents in 1850 were willing to support five or six kids with a negative financial return, why aren’t we?