« Predictably Irrational, by Dan Ariely | Main | Squaring the circle on trade »
Is divorce bad for the children?
Rereading Tim Harford's chapter three, this question runs through my mind. Numerous studies correlate divorce with subpar outcomes for the children, though Justin Wolfers once told me he was not convinced these studies had proved a causal relationship. Perhaps divorce-prone families have other dysfunctionalities which correlate with the kids having later problems in life and that the divorce is not causing those problems.
My wondering is more fundamental. Does a marginal (expected) increase in divorce increase or decrease the number of children who end up being born? On one hand the prospect of divorce may cause some people to limit the number of children they have. On the other hand, there is a surplus of women on the marriage market. Divorce, followed by male remarriage or at least siring, tends to increase the total number of children. I suspect this latter effect predominates. If divorce is unexpected, this latter effect almost certainly predominates.
If divorce causes more children to be brought into the world, it is hard for me to believe that divorce is bad for "the children" overall. It's better to be born, at least for most kids. You might argue that "children existing now" have a special moral privilege over "children in the abstract." Sometimes, yes (we don't value human life at replacement cost), but if we are asking "will divorce be good for children thirty years from now" currently they are all "children in the abstract."
I believe this defense of divorce is consistent with Tim's overall take. Of course a person who fears overpopulation might see this as additional reason to oppose divorce or make it more costly.
Posted by Tyler Cowen on February 3, 2008 at 09:10 PM in Philosophy | Permalink
Comments
Wait . . . Aren't you giving positive ontological status to yet-to-be-conceived children? Those 'children in the abstract' aren't anything. They don't have any sort of existence; so it seems wrong to claim that 'it's better to be born [than not].' There is no alternative to being born since there's nothing [a non-child] to which an alternative can refer.
Posted by: Jay Kelly at Feb 3, 2008 9:27:26 PM
If divorce causes more children to be brought into the world, it is hard for me to believe that divorce is bad for "the children" overall. It's better to be born, at least for most kids.
Old Pro-Life: Life begins at conception.
New Pro-Life: Life begins before conception.
Posted by: Jason Malloy at Feb 3, 2008 9:36:21 PM
High divorce rates raises the risk adjusted cost of having children for women since unmarried women with children often fall into poverty. Since women can control their fertility, I think is more likely that marginal increase in divorce decrease the number of births.
Posted by: joan at Feb 3, 2008 9:49:20 PM
Dr. Cowen's flirtation with the Guf aside, I can think of several lines of evidence that suggest divorce does not cause the negative outcomes that are associated with children of divorce. (that is family dissolution does not create behavioral problems)
For one, children who lose their father to an accident do not experience these divorce problems.
Divorce more likely helps children by getting them away from the stressful spectacle of a dysfunctional marriage:
"On average, children from single-parent families have more substance abuse problems than children of two-parent families. But levels of substance abuse are even higher among children of two-parent families who have a poor relationship with their father or experience high conflict between their married parents."
Posted by: Jason Malloy at Feb 3, 2008 9:59:14 PM
It's better to be born, at least for most kids.
Tyler, where's the evidence for this?
Bryan Caplan made a similar claim on his blog several months ago. His only support for the claim was that most people never choose to take their lives. That's true, but it doesn't seem very relevant.
Posted by: at Feb 3, 2008 10:24:05 PM
You can carp about the future and potential people all you want. If you agree that, thirty years from now, a world with five billion people is better than world with five people (as indeed you must agree), the rest of my argument follows.
Posted by: Tyler Cowen at Feb 3, 2008 10:40:12 PM
It's better to be born, at least for most kids.
This blog disagrees, as does this one. Both of them are engaged in a discussion at my blog here.
Posted by: TGGP at Feb 3, 2008 10:52:33 PM
My situation is interesting. I married a divorced woman and we had two more kids (in addition to her first). Her ex-husband remarried and had one additional child. Not sure if this increased the net number of children born or not.
Posted by: Fungible at Feb 3, 2008 11:22:13 PM
a world with five billion people is better than world with five people (as indeed you must agree),
i don't see why he must agree. if a 80 year old person with no family connections or friends younger than he is probably wouldn't care what the world is like in 30 years.
Posted by: a person at Feb 3, 2008 11:43:55 PM
"If you agree that, thirty years from now, a world with five billion people is better than world with five people (as indeed you must agree), the rest of my argument follows."
Isn't this blog called Marginal Revolution? Wouldn't the relevant question be whether the world would be better to have 5 billion versus 5 billion and one? The answer to that one is far from obvious.
Besides, whether or not divorce is bad for the kids is an empirical question that I thought was settled in 2001 by the Hetherington study.
Posted by: M. Hodak at Feb 3, 2008 11:45:50 PM
I asked the same question in Harford's comments after he posted the excerpts. Cowen's answer is bolstered, at least anecdotally for me from my own experience and that of people I've known. I don't have the stats to back it up, but it seems to me that 2nd marriages have a much greater chance of lasting and the children that result get more stable, Ozzie and Harriet style upbringings. From what I know of the Daddy-O I haven't seen in 20+, he did all right by the 2nd litter. This seems to be pretty common.
So not only are there more kiddos, these kiddos-in-the-abstract have a pretty good shot. And even the abandoned ones tend to get their bearings eventually, though they may flail around for a decade or two first (and wonder how they parlayed high-g into librarianship and wonder just how much they hurt their livers during all the flailing).
My personal suspicion is that divorces cause real harm, but less than bad marriages. And I tip the hat to my step father, Cowen, and all the others that raise the Dodo eggs--though I would never do it myself in a million years.
Posted by: burger flipper at Feb 4, 2008 12:01:49 AM
I think the point can be made without the apparent silliness of "better to be born than not"-argument.
Let us say there are two kinds of families, divorce-prone and regular. Children in DP-families are worse off than regulars. Then, if there is some number of new regular marriages (arguably less than 2) as a result of the divorce, then any children born in those will raise the average childs well being.
This of course assuming that the new marriages will not be DP.
Posted by: Tiedemies at Feb 4, 2008 12:59:49 AM
"It's better to be born, at least for most kids."
I use the same argument when discussing animal rights. More animals are born and living (relatively) healthy (albeit short) lives because of the meat industry.
I think one ought to specify what is to be maximized for this being a relevant discussion. Total utility for "the children" (which, undoutedly grows with each new child ) or average/median utility (which doesn't necessarily grow with each birth.)
Posted by: Robert at Feb 4, 2008 2:15:11 AM
It's worse to be born as a slave than not to be born.
It's worse to be born as a polio victim than not to be born.
It's worse to be born as a dyslexic than not to be born.
It's worse for anyone who experiences any pain at all (including death) to be born than not to be born.
Any effort that is spent on those you caused to be born should have been spent on relieving the suffering of those unfortunate enough to already have been born.
Posted by: Noumenon at Feb 4, 2008 2:38:03 AM
Dr Cowen,
Did you ever develop a serviceable theory on why we don't value human life at replacement cost?
While I'm sure you no longer have to worry about out of left field interview questions, it seems like just the kind of issue to which you may have given some thought.
My best guess, based on 2 minutes of intense contemplation, is that there is a kind of halo effect.
For obvious reasons I value my own life at very much more than replacement value. Likewise I value the lives of my family and friends very highly.
While this doesn't directly lead to the placement of a high value on the life of a random stranger, it isn't hard to imagine that a random stranger has friends and family who value him at much greater than replacement cost. If we believe that other people consider something to be valuable that belief affects our perception of its value.
It seems conceivable that our valuation of people that we are close to leads us to extrapolate higher than replacement cost valuations for people we have never even met. A non-economist might refer to this as empathy.
Posted by: Deepish Thinker at Feb 4, 2008 2:46:56 AM
It's better to not have been married.
"Don't gamble; take all your savings and buy some good stock and hold it till it goes up, then sell it. If it don't go up, don't buy it." Will Rogers
If you are going to get divorced, don't get married.
Why start from the assumption of a crappy marriage? But even starting from there and assuming the only options are (A) stay in crappy marriage or (B) divorce crappy marriage; Go ask a kid whose parents are divorced whether they think divorce is good for the kids.
I'm always fascinated by people who get divorced, then get remarried. What has changed? It's like football coaches. Why would you hire someone who failed elsewhere?
Posted by: Andrew at Feb 4, 2008 4:32:50 AM
I think we do value life at replacement cost. Select at random your favorite loved one. What would be the cost to replace them? Just because the law doesn't recognize it this way doesn't mean it accounts for how we think and feel about it. I think the value the law places on pets killed by negligence is replacement value. If you have a pet, you will understand how ridiculous that is.
Many people do commit suicide. The ones who don't aren't necessarily not doing it totally because they feel they are better off. There is a component of a sense of commitment to those who care about you. I once heard a therapist say that if you get someone you suspect a suicide risk to promise to call you before they do it, they will almost always fulfill that promise. If you can get them to sign a contract it's even better. People have a tendency to fulfill their contracts and this same concept, in an informal way, keeps a lot of people from taking their lives. How many? Who could know?
"It's better to be born" Yes, for the ones who are born!
To take it back to Levitt and Dubner, mothers already make decisions about whether they can properly raise a child. Most make that decision before conception, and many afterwards. Of course, some are mistaken or never even think. So, to reason that the next marginal child is better off being born than not because all the ones who are born (based on the mother's decision that they are better off) begs the question. If mothers were perfect in their decision making and prioritize the well-being of the child, then the next child that would have been born is not better off, because the mother decided not to have the kid. Obviously, they aren't perfect and don't always prioritize the well-being of the child, but they tend in that direction.
Posted by: Andrew at Feb 4, 2008 5:38:05 AM
"You can carp about the future and potential people all you want. If you agree that, thirty years from now, a world with five billion people is better than world with five people (as indeed you must agree), the rest of my argument follows."
If you are putting words in other's mouths, please try a little harder to understand their thinking. I do think that the world would be a better place with fewer people. Not five of course because that would not be viable. I think that the world with one billion instead of five billion would be nicer, more livable, and more sustainable over the long haul. And personally I don't give a crap about the unborn.
Of course the problem with reducing the population is not the ridiculous claim on the part of the unborn. It is the very real and legitimate desire of born people to have children.
Posted by: RobbL at Feb 4, 2008 7:22:18 AM
All I can suggest to those of you who buy into the "better never to have been born" camp is to please get some therapy, prozac, quit the job you hate, or find out what you want to be doing that you aren't. Or emigrate from South Africa, for that matter.
There is a vast amount of good information about the way that mood and underlying outlook influences thought processes, and vice versa. You are a victim of that phenomenon, who has imbued these negative thought processes with the status of "rational beliefs".
Posted by: Matthew at Feb 4, 2008 8:03:14 AM
"It's worse to be born as a slave than not to be born.
It's worse to be born as a polio victim than not to be born.
It's worse to be born as a dyslexic than not to be born.
It's worse for anyone who experiences any pain at all (including death) to be born than not to be born."
WTF?!
http://www.independentliving.org/docs3/gill99.html
This is just one example because I am in a hurry. Disabled people rate their quality of life higher than able-bodied people, as a rule.
I know happy-to-be-alive people who are dyslexic, had polio, have chronic severe pain, etc. Ironically (but thankfully), I even know depressives who are (most days) glad to be alive.
And why did/do most slaves try to survive or get out of their situation, rather than kill themselves? Not saying it's a nice life, not saying it's just or anything...
But you are creating a false dilemma here. Or something a lot worse that I don't know the name for.
Posted by: natalia at Feb 4, 2008 9:37:12 AM
Oh, I meant to say this also (before that other disturbing comment caught my mind):
If the parents get divorced because one of them was abusive (especially towards the child/ren, although any abuse they lived around could be damaging to them) then divorce is better for the child/ren.
Just wanted to throw that in there.
Posted by: natalia at Feb 4, 2008 9:40:20 AM
For those interested, Pitt economist, Mark Hoekstra has a very interesting paper trying to identify causal relationships between divorce and child outcomes. He uses what feels almost like a regression discontinuity design to separate correlation from causality by looking at a sample of families in divorce court who went through with the divorce versus families who decided to stay together, since both families were a sample of self-selected divorcers. This may not be enough to actually identify the effect, since the decision to remain together is still potentially endogeneous, but nevertheless, it seems like it takes us further than most studies. Here's the abstract:
Previous research has identified the effects of parental divorce primarily by comparing the
outcomes of children whose parents divorced to those of children in traditional two-parent
families, conditional on observable characteristics. In contrast, this paper utilizes an exceptional
panel of child-level data containing observations of children whose parents filed for divorce but
did not subsequently divorce in addition to those of children whose parents did divorce. Using
this panel of child-level administrative data on reading/mathematics composite test scores and
disciplinary records for a large Florida school district, this study yields several notable findings.
First, while the data confirm that students whose parents divorced have lower test scores and
more disciplinary problems afterward relative to children whose parents never filed for divorce,
the same is true of children whose parents filed for divorce but dismissed the case. This suggests
that the differences observed in the classic comparison of children of divorce to children of two-
parent families are due to the factors that caused the parents to divorce and not to the legal
dissolution of marriage itself. Second, when comparing children whose parents divorced to those
from the dismissal group, I find that children who experience a parental divorce score a
statistically insignificant 5 points higher than those whose parents went to the brink of divorce but
remained married, the standard error of which allows me to rule out meaningful negative effects
of divorce on test scores. In addition, while I find no evidence of a lasting effect of parental
divorce on disciplinary problems 4 years afterward, I present suggestive but inconclusive
evidence that girls achieve at higher levels as a result of divorce than they would have had their
parents remained married.
Posted by: jason voorhees at Feb 4, 2008 9:48:06 AM
Matthew, what if an anti-natalist suggested that you take a drug that would cause you to agree with them? It takes the desirability of their position as a given. If you want to confront the arguments of anti-natalists, do so rather than simply asserting there must be something wrong with them, which can even more aptly be turned around on you given that the depressed (though I do not know if any anti-natalists actually are depressed) tend to have more accurate beliefs about everything except the probability of their own recovery from depression.
Posted by: TGGP at Feb 4, 2008 10:20:16 AM
Singular, albeit defining, moments in one's life do not necessarily determine the final product of that life. Divorce included. Fatalism aside, one dedicated parent can save a child from whatever statistical inevitability a divorce has "pre-determined."
Posted by: Chicagoan at Feb 4, 2008 10:35:37 AM
Eyeballing this list leads me to conclude that divorce rates are not a critical distinguishing factor in determining the rate of total fertility.
The differences between Mali and Hong Kong are legion and the explanatory power of analyzing divorce should prove limited.
Posted by: John Sterling at Feb 4, 2008 10:38:40 AM
