Bryan Caplan  

The Weird Reason to Have More Kids

U.S. Health Care Costs... Morning Mark...
Think of a trait that brings people together.  It could be jokiness, religiosity, libertarianism, love of books, or fascination with role-playing games - or seriousness, impiety, statism, hatred of books, or contempt for role-playing games.  Take your pick. 

Now suppose that the parent-child correlation on the trait you picked is exactly zero.  Then no matter what you're like, you should expect your kids to be at the 50th percentile.  If you're normal, that's a pretty good deal; at least on average, your kids will be just like you.  But the weirder you are, the less your kids will typically resemble you.  Even if you're at the 95th, 99th or 99.99th percentile, you can expect your kids to be perfectly average.  In a world of zero parent-child correlation, weird people have little in common with their children.

In the real world, of course, parent-child correlations almost always exceed zero, and are often substantial.  This doesn't boost the similarity between normal people and their children; no matter what the parent-child correlation is, parents at the 50th percentile typically have children at the 50th percentile.  But a positive parent-child correlation does boost the similarity between weirdos and their children - and the weirder you are, the bigger the boost.  Take a look:


Parent-Child Correlation



r= 0


























Notice that regardless of the value of r, normal people can expect to be like their kids.  But that's not saying much, because normal people can expect to be like any random person they meet!  The story's very different for weirdos.  By definition, weirdos never have much in common with random strangers.  With a zero parent-child correlation, weirdos will feel equally alienated from their children.  As the parent-child correlation rises, however, weirdos' incompatibility with strangers stays the same, but their expected compatibility with their children gets stronger and stronger.

Now let's look at these facts like a mad economist.  There are two ways to surround yourself with people like you.  One is to meet them; the other is to make them.  If you're average, meeting people like yourself is easy; people like you are everywhere.  If you're weird, though, meeting people like yourself is hard; people like you are few and far between.  But fortunately, as the parent-child correlation rises, weirdos' odds of making people like themselves get better and better.  This is especially true if the parent-child correlation largely reflects nature rather than nurture, because you won't have to ride your kids to emulate you; they'll do it on their own initiative.

You might object that meeting like-minded friends is always easier than making like-minded children, no matter how high the parent-child correlation gets.  If you're only looking for casual relationships, that's probably true.  But if you're looking for deep, time-intensive, life-long connection, blood really is thicker than water.

The lesson: As your weirdness increases, so does your incentive to have kids.  If you like football and American Idol, you're never really alone.  You don't need to build a Xanadu for yourself.  But if you're a lonely misfit, oddball, freak, or weirdo, then find a like-minded spouse and make new life together.  Let the normals laugh at you.  You'll have each other.

P.S. I'm speaking at Western Carolina University on Friday at 2:30 in the Forsyth Auditorium.  If you're there, please introduce yourself. :-)

Comments and Sharing

COMMENTS (26 to date)
Bob Murphy writes:

Bryan (or someone else who wants to speak for him), are you making an assumption that the traits are distributed in a bell curve, or at least, that they are clustered in the middle?

E.g. if there were a uniform distribution over the finite spectrum of a given trait, then it's not really true that a 50 percentile parent could expect to have a kid "just like him" except in the sense that I "expect" to roll a 3.5 on a die, right?

Phil writes:

But what if the kid turns out to be not like you at all? If you're weird in the 99th percentile, and the kid is average, he may grow up to dislike you, or at least not really bond with you. And a child that doesn't get you is a lot harder to take than a a stranger who doesn't get you.

Cuts both ways, doesn't it?

Zaxecivobuny writes:

This is really bad reasoning. First of all, it is pretty messed up to have kids so that you have someone to hang out with. That puts a lot of pressure on the kids to conform to your expectations, which is already pretty unhealthy. Secondly, thinking like an economist means thinking about incentives. Some people are predisposed to have lots of children, but others see the implicit and explicit costs as not worth the benefits (or the wait time until they are more fun to hang out with rather than more of a burden to maintain). Moreover, the internet has drastically reduced transaction costs in finding people who have similar outlier preferences to oneself. You don't have to rely on the (very poor) process of screening strangers with a low probability of being star trek fans who also like care bears, you can just search facebook for that overlap of interests, for example.

Ilya Somin writes:

It might still be cheaper and easier to find additional close friends with similar interests than to have additional kids. With close friends, you can still have "deep, time-intensive, life-long connection," and you don't have to pay for their upkeep, etc.

Moreover, generational differences may reduce the degree of commonality of interests between parents and children, whereas your close friends are probably going to be contemporaries.

Don't get me wrong. I think there is a lot of merit to your overall case for having more kids. But I think this particular argument is overstated.

Hyena writes:
If you're weird, though, meeting people like yourself is hard; people like you are few and far between.

He says unironically over a medium that's been bringing extremely weird people together since 1990.

I think this argument works better as a reason not to avoid having children, provided you like yourself decently enough. They will, after all, probably turn out to be someone you like....

Jeffrey Horn writes:

Bryan: What about reversion to the mean? For normally distributed variables, like height, being in the tails makes it nearly certain your children will be closer to the mean than you are (despite a high parent-child correlation for height). Would the correct way to reconcile these concepts be that, if you're weird, your children will be weird, but almost certainly less weird than you?

I know some people would take that interpretation as good news, as well.

Pandaemoni writes:

Then again, what is one of your traits is a desire to be as happy as possible? Every study I have ever seen has suggested that parents are less happy than otherwise similarly situated non-parents (even after the children are grown and have left the home, the difference persists). So it seems to me that if you are sufficiently desperate for a reason to have children, that this analysis is the hope you cling to, then you should consider the more basic question of whether you'd be happier without children. (In fact, I really think everyone should consider that basic question, but recognize that biological drives tend to trump Reason in this area of life.)

One positive note, though, is that the studies of the impact of children on parental happiness do not focus on the oddballs. Presumably the studies' results reflect the "average", so they apply more strongly to the people who are predominantly normal, than they do to the weirdos.

Assuming you have a number of positive traits in which you are an outlier and a limited number of negative traits, then it may be that you can expect to overcome the reduction in happiness that the studies otherwise suggest (since your kids will also be expected to evince heightened levels of those traits, and so raising your kids may have either fewer moments of worry and hassle, more moments of joy, or heightened levels of joy in those moments).

Floccina writes:


I would like to read Bryan on the risks of child bearing.

I did not have more children because having children is risky. You could have a severely retarded, handicapped or criminal child. You could, if you are the mother die in child birth.

Peter Finch writes:


> I would like to read Bryan on the risks of child bearing.

I would too.

I am mostly convinced by Bryan's pro-kid arguments. Heck, we've got three-going-on-four.

But anecdotal evidence makes me suspect that the small negative happiness associated with children is clustered in Severe Adverse Event people. That is, I suspect children make people happy, except when they die of cancer at age four, or have a survivable but horribly debilitating defect. The odds of an SAE go up with each child you add to your brood. I'd love to see numbers on the distribution of happiness among parents. Is the median near the mean?

On topic, despite my best efforts, I've found it hard to maintain a "deep, time-intensive, life-long connection" with friends in the modern world where people move for work and school. It's been much easier with my siblings. Maybe if you live in a small town and will probably die in the small town. But we mostly don't live in isolated tribes anymore; friendship might be less useful today.

Peter Finch writes:

The SAE reasoning would also explain why the small negative happiness effect persists after children move out: a devastating event is still devastating, even when your child is 28. And why first children are associated with more unhappiness: people discover their predisposition to heritable problems with their first child, and then cease having further children.

agnostic writes:

Humans do not reproduce asexually. Therefore the "You" column in the table and the analysis is the "mid-parent" value, or the average between the mother and father's values in SD's.

The strength of the effect you point to is only that big if your mate is also as weird as you are. The closer they are to normal -- including the case where your mate is on the other side of the average -- the more diluted the correlation between the weirdo parent and the child.

So to get a powerful effect, you already have to search out a weirdo in real life -- your mate -- and then wait a long time to see your kid mature. For friends, you only have to do the first part, and again the effect won't be diluted by regression to the mean of children.

Let's make a quick reality check -- are weirdos more likely to seek out weirdo friends or have kids to provide them with weirdo companions? Everywhere and since time immemorial, it's the former. That's the stable equilibrium, so that's the right choice for them to make. Whenever some incredibly wrong-looking proposal is made to benefit some group, you should always ask why they've never thought of it before.

Peter Finch writes:


Personal observation: people with weird kids often have more kids so that their kids can have close friends who are like them. "Weird" is poor word choice, since it has a negative connotation. You see this in people with mild disabilities like dwarfism, or in people who are culturally isolated like immigrants. People with weird traits certainly mate disproportionately with others who have that trait.

I didn't think the behavior Bryan described was unusual at all. I'm surprised to find people think it is.

dave smith writes:

This is easily tested. Do weirdos have larger families? Well, maybe not easily tested...lots of endogeneity!?

Miguel Madeira writes:

There is an hidden premise in Caplan's post: that there is an advantage in "surround yourself with people like you".

agnostic writes:

Peter, Bryan is trying to make a case for the superiority of making weirdos compared to searching for them. You're describing assortative mating, and that's prevalent enough. What I'm saying is that this proves that searching is better than making weirdos -- in order to get the full benefit of Bryan's parent-offspring correlations, you have to search for a mate who is as weird as you are, given sexual reproduction.

Also, if you want to hang out with someone who's a weirdo like you are in all traits, that is far easier when you're searching for rather than making more weirdos. Namely, find a number of friends who among all of them will allow you to hang out with someone who's as weird as you are in all relevant traits.

Trying to squeeze all of that into one person would mean cloning yourself phenotypically (not genetically, though that'd be good). Some weirdnesses do highly correlate, so it's no extra effort, like finding someone who's into both NASCAR and hunting. But the suite of traits is not, making it much more costly to search for a single weirdo like you in all dimensions, with whom to mate and produce another one of you.

Peter Finch writes:


I realize what I described is slightly different than what Bryan described, and you have a point. It is easier to find people who are like you than to make them. Although it's much harder to keep those found people. And, as Bryan would admit, it's only possible to make children similar to you in certain respects.

But people do things that are very similar to what Bryan describes: they find mates that are similarly weird, and they choose to have more children specifically because of their weirdness. Obviously, I don't have numbers to back this up - but hang around families with disabled people and you'll hear this kind of thinking.

It's got to be a small effect, and it's probably useless if your weirdness is "liking role-playing games". And children are probably overrated as companions, although that's a common reason for having them, too. They are great companions for siblings, as I pointed out in my earlier post.

Groucho Marx writes:

I wouldn't want to hang out with anyone as weird as myself.

Steve Sailer writes:

Say you are weird in that you are at the 95th percentile in procrastinating. You marry someone with the same proclivity toward procrastination. Your children thus have both nature and nurture making them tend toward procrastination. Is this a good combination?

Peter Finch writes:

> Say you are weird in that you are at the 95th
> percentile in procrastinating. You marry someone
> with the same proclivity toward procrastination.
> Your children thus have both nature and nurture
> making them tend toward procrastination. Is this
> a good combination?

Good for who?

Some negative traits are less negative if there are others around who have them. Deafness comes to mind.

JayJay writes:

Hilarious. This has really given me added incentive to have kids. I thought having kids would alienate me even more from people I have little in common with due to my inevitably weird parenting style, but I'll be replacing them with a brood that is more likely to be able to relate to me anyways. Doesn't help that my partner is also weird. How our kids will ever be able to communicate with the rest of humanity is beyond me.

David Friedman writes:

"This doesn't boost the similarity between normal people and their children; no matter what the parent-child correlation is, parents at the 50th percentile typically have children at the 50th percentile."

That's wrong.

The parent at the 50th percentile will on average have a child at the 50th percentile, with or without a correlation. But without a correlation, the child's characteristic will be on a distribution centered at the 50th percentile but with weight from 0th to 99th. With a correlation, the child's characteristic will be on a similar distribution but one more heavily weighted towards the center--how much more depending on the correlation. So the correlation increases the chance of parent/child similarity for the parent at the 50th percentile as well as the one at the 99th percentile.

David Friedman writes:

So far as Bryan's central point, my experience supports it. All three of my children are much more "my sort of people" than most other people I know. That isn't, by itself, an adequate reason to have children, but it is one more benefit to be counted in making the decision.

Psychohistorian writes:

This actually seems like extremely strong evidence against the theory that parents make no difference; at least, it reminds me of such evidence.

If you are a member of a weird religious cult, the odds that your children will also be members of that cult are staggeringly high compared to random strangers. If you enjoy any particular activity or set of activities that can be shared with kids, there is likely to be a much-better-than-random chance that it will be shared by your kids. Huge-picture life outcomes as measures by economists may not be clearly transferable, but things like religion and hobbies almost certainly are.

Advocate writes:

[Comment removed for supplying false email address. Email the to request restoring this comment. A valid email address is required to post comments on EconLog.--Econlib Ed.]

Ed writes:

Did someone say "unironical"?
Wonderful post -- and no better or worse than any other reason to have or not have kids IMHO, just more fun to read -- many of the comments too.

Pandaemoni writes:

Forgive the very late post that likely no one will ever see.

Regarding the recent charging of Professor David Epstein of Columbia with having had a consensual sexual relationship with his (adult) struck me as curious that that would be likely a very efficient way to have kids with the maximum possible similar oddness to oneself.

Link to the story.

Actually just having read it now, I am presently struggling with how I should condemn a consensual adult incestuous relations (not just parent child, but siblings, uncle-niece and aunt-nephew, cousins, etc.). I know that I *want* to condemn them, but I wonder if proceeding from the emotionally based premise will lead me to an answer.

Comments for this entry have been closed
Return to top