Susan Mayer's
What Money Can't Buy concludes by tossing out a fun fact I've often heard repeated. (I even repeated it myself once in
an exchange with Charles Murray).
Both low income and single parenthood may in fact be correlated with poor outcomes for children because they are proxies for unmeasured parental characteristics. This suspicion is bolstered by the well-established finding that when single parenthood is a by-product of death rather than divorce or failure to marry, children do about as well as children living with two parents who have comparable incomes. (McLanahan and Sandefur 1994)
So I decided to check out McLanahan and Sandefur (1994), entitled
Growing Up With a Single Parent. Mayer's claim checks out, but the evidence was thinner than I hoped. Here's the key figure:
The behavioral genetic story, of course, is that in modern societies, premature death is usually a random fluke. As a result, a parent's premature death doesn't tell us much about his children's life outcomes. Divorce and extramarital births, in contrast, are largely products of parental behavior, which in turn substantially
stems from hereditary traits like IQ, conscientiousness, neuroticism, etc.
As always, other stories are possible. But given
everything else we know about nature and nurture, not very plausible.
SHARE
Use "heritable" rather than "hereditary" -- or people will accuse you of being a 100% genetic determinist.
Death often comes with life insurance, a big surge of financial resources. Divorce often reduces total spending on childrearing.
Any given person's chance of premature death may be pretty random in modern society. But across large populations, I'd expect people with a tendency to engage in risk-seeking behaviours to be more likely to die prematurely than the conscientious. And if there's a genetic influence on engaging in risk-seeking behaviours, then that could account for the slightly higher risks in the widow relative to divorce situation.
Does the study address when the death occurred?
If you compare children whose father died when the kid was 2 vs. children whose father died when the kid was 7, is there a measurable difference?
You'd be dealing with unfortunately small sample sizes, but it could give some additional measurable influence of parenting.
The same method could also be applied to divorce.
[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]
I want to second the point about life insurance. Even without any insurance, your partner divorcing you is about the same financial burden as them dying (possibly moreso depending on the circumstances, such as raiding of joint bank accounts and giant attorney's fees). But death often comes with a large life insurance payout (and any giant medical bills can't exceed the size of the estate).
If family finances matter to outcomes, then divorce would be expected to be much worse.
> If family finances matter to outcomes, then
> divorce would be expected to be much worse.
While I think there's a lot of truth to what you say, divorce doesn't stop one of the parents from working forever. Death does. That's a significant financial cost.
And just how common are large life insurance payouts for young parents?
Divorce has a more substantial impact on a single-parent's finances, sure. But should that impact not be part of the consideration as to why it is demonstrative of undesirable heritable traits in the first place?
The critique of mothers bearing children out of wedlock is, in part, that they are irresponsible for not finding some measure of financial security before childbirth. Divorcees, if we grant the negative financial impacts alleged, are making the exact same choice at a different time.
Mayer's claim checks out, but the evidence was thinner than I hoped.
Expanding on Tracy W's point, the top 5 causes of death between ages 15 and 34 are traffic accidents (13,000), suicide, poisonings, violence (9-10,000 each), and other injuries (2,000).
All of these correlate with risky or irresponsible behavior. Looking at death by illness would probably give you the stronger evidence you are looking for.
There are a number of different kinds of studies that all converge on this. For example, lack of harm in father absence from extended military service. Another example is the resulting similarity of children adopted by single mothers with children adopted by couples. And—my personal favorite—the evidence of harm from father presence in subgroups where father absence is popularly taken for granted as an important cause (rather than a correlate) of dysfunction. Single motherhood appears to be a more beneficial arrangement for children if their fathers are losers.
Completely I share your opinion. In it something is also to me this idea is pleasant, I completely with you agree.