Do parents matter? One way to find out is to compare how children who suffer the death of a parent fare in later life compared to those who don’t. This new paper does just this for Swedish kids, with results that might be surprising:
Parental death has surprisingly small average effects on cognitive outcomes, despite representing a traumatic shock. Given the size of our dataset, we can rule out zero effects, but our preferred estimates represent a loss of a couple of months of schooling.
The impact on subsequent earnings is also small - 6-7% for boys and 1-2% for girls.
Curiously, it doesn’t make much difference whether it is the mother or father who dies. Nor does a parent’s death when the child is young have a larger effect. If anything, the opposite - for boys, the death of a father before the age of 10 has no significant impact upon education, IQ or earnings.
Quite why this should be so is not so clear. It could be that the death of one parent causes the surviving one to try harder. In The Talent Code, Daniel Coyle suggests another possibility. The loss of a parent, he says, sends out a “primal cue” - that one is not safe, and had better work hard to make something of oneself. This, he says, unleashes the energy that is necessary for success: he lists a huge number of eminent people (Darwin, Newton, Lincoln, Gandhi, Dostoyevsky etc) who lost a parent when young.
This might corroborate Bryan Caplan’s view that “the long-run effects of parenting on children's outcomes are much smaller than they look.” It’s also consistent with Ian Walker and Yu Zhu’s finding that it is a father’s income that matters for his child’s success, rather than his presence.
That said, this is also consistent with the possibility that the children of single parents generally do worse than those of dual parents. This, though, would be because the sort of deadbeat dad who abandons his children is also the sort who transmits “bad genes” (impatience?) to them, rather than because he deprives them of valuable parenting skills.
Could it be that the opinion that parenting matters - within reasonable limits - owes more to the egocentric bias than to the evidence?