Bryan Caplan  

Savings, Genes, and Fade-Out

Joffe on LIBOR... A Paragraph to Ponder...
Parenting often has large effects on the young.  Parents do stuff, their kids respond, and observers conclude that parenting is very important.  You need twin and adoption methods to uncover the crucial caveat: these parenting effects usually fade-out.  Kids aren't like clay that parents mold for life; they're like flexible plastic than responds to pressure but returns to its original shape when the pressure is released. 

Latest example: savings behavior.  Cronqvist and Siegel's "The Origin of Savings Behavior" compares identical and fraternal Swedish twins to measure the relative effects of nature and nurture over the life cycle.  Parenting has a big effect on savings behavior in your twenties, but rapidly vanishes:
[P]arenting explains about 40 to 50 percent of the variation in savings rates for the youngest individuals in our sample (20-25 year olds), but this effect decays significantly and attains zero by middle age. That is, while parents seem to strongly affect their children's savings behavior early on in life, the effect disappears over time as their children gain their own individual-specific life experiences.

Our interpretation of this evidence is that social transmission from parents to their children affects children's savings behavior early on in life, but unlike genetic effects, parenting does not have a lifelong impact on an individual's savings behavior. These results are broadly consistent with research in behavioral genetics which has found a significant effect of the common family environment in early ages on, e.g., personality, but also shown that such effects approach zero in adulthood...
Cool graph:

You might think that, due to the magic of compounding, parents' strong effect on early savings would translate into a strong effect on retirement wealth.  Not so:
We find that about 39 percent of the cross-sectional variation in wealth accumulated up to retirement is explained by genes. The effect of the common family environment and upbringing, which by model construction also reflects wealth inherited from parents, explain seven percent of the cross-sectional variation.
What's the underlying mechanism?  Cronqvist and Siegel's other results point to lasting nature effects - and transient nurture effects - on time preference, patience, and/or self-control:
We find a significantly positive correlation between an individual's savings rate and income growth, consistent with the common prediction that patient individuals experience higher income growth. Importantly, for a subset of individuals, data on education (number of years), smoking (number of cigarettes smoked per day), and body mass index (BMI) (weight relative to squared height) are available from the STR's interviews. We find no statistically significant correlation between education and savings behavior, but... those who save less are found to smoke more and are more likely to be obese...
As usual, neither nature nor nurture come close to a full explanation of savings, income growth, smoking, or obesity.  The effect of "non-shared environment" - also known as "none of the above" - is massive.  If free will is a key determinant of economic success, this is precisely what we should expect.

HT: Rob Wiblin

Comments and Sharing

COMMENTS (9 to date)

This might mean nurture was far more important when life expectancies were lower. In other words, "tiger mothering" used to make sense.

Steve Sailer writes:

Okay, so parenting is real important up through the age when people get started on careers and are picking out mates. That makes parenting sound rather important.

Miraj Patel writes:

There has been a large body of work on a variety of topics looking at the effects of environment vs. genetics on children and how it impacts them throughout life in psychology and genetics research. At the end of the day, I think the data is strong enough to conclude that they are both substantial parts of who a person becomes, how they think, how they act, and how they see the world. Parenting matters in the short and long run. So do genes.

Joe Cushing writes:

This feels like a very long post for what seems obvious from the title and the graph. I can't get myself to read it. Is there anything I'm missing?

JayMan writes:

Free will, of course, does not actually exist. Non-genetic sources of behavior (by definition "environmental"), don't actually restore a place for free will. Rather, they replace genetic determinism with environmental determinism.

On the most fundamental level, all actions—indeed, all events in the universe—are either strictly predetermined by the physical state of the particles that comprise them—in which case, we are not "responsible" for them, or are caused by random events (say, quantum effects—which is almost certainly the case in actuality)—in which case we are not responsible for them. The putative "free will" simply cannot exist because of this reality.

All actions, in living and non-living systems, have causes. That's a basic consequence of a physical universe dictated by physical forces. The individual human is not some agent divorced from this physical reality but a construct within it. The behavior of the individual is determined—even if, ultimately, only probabilistically—by the physical forces acting upon the particles that comprise the individual.

On a higher level, closer to the level the causes of human behavioral traits with respect to differences in those traits between individuals, there still can be no place for free will. It seems to me that people often say "free will" to mean "mutability". That is, that one can "change" oneself—or be "changed" by the right conditioning or treatment; the alcoholic—whether an alcoholic by genes or by the result of experiences—can "choose" to not be one, or be "made" to not be one. This view is also rather naive. Sure, people do seem to lick recurring behavioral attributes they posses (that is, "traits"), but the reality is that such cases are themselves probably best described as a complex interaction of such traits (say between high determination and the alcoholism—a tortured soul be such a person ), and, of course, our good friend, the aforementioned randomness. Personality—and behavioral traits overall—appear to be very stable characteristics of human psychology. We are, for better or worse, who we are. This is precisely what one would expect if behavior was the result of actions in the brain, actions that play out depending on the brain's physical structure and programming encoded in that structure. This is the reason that depression is so hard to treat, that the War on Drugs continues to rage, that the obesity epidemic continues out-of-control, and that criminals such as pedophiles resist the formidable deterrents placed before them.

And as for that "unshared environmental component", it is, as Judith Harris freely admits, in actuality just the "unexplained variance" in behavioral geneticists equations. It is thought that this is where the "environment" can work its influence. Judith Harris posited that peers have an effect here, which was described in a well-reasoned manner in No Two Alike, but I'm not so sure the peer impact is as big as we might think. As far as I know, there isn't a whole lot of evidence for the peer effect. That children who hang together tend to be alike may be entirely due to self-sorting. My own suspicion is that the peer effect operates on some traits more than others, with self-esteem being an example where peer experiences likely do have impact. However, with many other traits, a substantial amount—or perhaps all—of the "unexplained variance" could be due to developmental noise. This is the force that leads identical twins to have different fingerprints, and perhaps also to have different sexual orientation. If so, many aspects of behavior, while only partially genetic in origin, could still none the less be wholly "biological".

Part the theory Harris' posed in No Two Alike is that individuals develop their traits—in a way such that MZ twins grow distinct—because we are designed to differentiate ourselves from all other individuals. One might see an opportunity to insert free will here, but what we know about how brains work would lead one to question this view. Likely this process operates largely outside our "control". As with all other things, this process probably happens largely "automatically", with our conscious "selves" then painting a story that gives us the appearance of being in control.

These aspects of humanity are why classical Greek tragedy remains so enduring. It is an accurate reflection of the human condition.

As for the finding of the original post, I'm not surprised. This is consistent with all the other evidence that our personality traits are not shaped by our parents. In reference to Steve Sailer's comment, that parenting cannot shape who we are doesn't mean (nor did anyone ever claim) that parenting can't shape where we end up in life. All else being equal, one's life is going to be a whole lot different if one is the child of wealthy parents vs being the child of paupers, even if one's personality is not impacted. Parents, in a fashion, are one of the providers of opportunity, and opportunity is a factor that is so key.

I think the key takeaway is that parents who are concerned about their children's future should stop fretting about shaping the way they behave or the values they possess and more concerned about giving them tips and tricks and using their pull to open doors for them. The former two are quite distinct from the latter two.

Murali Vajapeyam writes:

Maybe in a society where people only lived to their 40's, and lived most of the time with their parents, the influence of parenting would last for most of the children's lives? It is possible that the whole "parenting has no long-run effect" is rather a byproduct of more recent societal developments, i.e. people living longer and leaving home. Once that started to happen with societies, nature and genes took over the long run...

JayMan writes:
It is possible that the whole "parenting has no long-run effect" is rather a byproduct of more recent societal developments, i.e. people living longer and leaving home.

But if it's true now, it had to always have been true, or at least had been true for a long time. Such a fundamental change in how human personality develops couldn't have evolved so quickly.

Indeed, the lack of a long-term effect of parental influence is a key feature of Robert Trivers' theory of parent-offspring conflict, and he predicted what behavioral geneticists would later find all the way back in 1974, which he beautifully described (see the end of pg 10 to pg 11).

Murali Vajapeyam writes:

JayMan: I think we are saying the same thing. Trivers' writing in 1974, as well as pretty much all twin studies, are recent. Any research in the last 100 years would fit within my time frame of recent, resulting in the unanimous conclusion from research that "parenting has no long term effect". But if twin studies had been carried, say, in the 18th century, we might have seen different conclusions.

And I am not saying that how genes operate to develop personality has changed -- I agree that genes/humans/evolution are largely unchanged over the last 1,000 years (probably way longer). But the phenotype, i.e. the neutrality of parenting in the long-run, is a recent phenomenon, and was likely not observable for most of modern human history -- it is a byproduct of thousand-year-old genes operating in a novel environment never before seen.

So, Bryan Caplan is right that parenting is like bending a tulip. But what I am saying is that for most of human history, the bending was there for most of the tulip's life, so the effects of parenting were close to life-long.

Maoz writes:

I don't find the distinction between parenting and "non-shared environment" particularly useful in terms of explanatory power. Parents have a role in determining non-shared environment. If my nagging gets my child to go into Honors English instead of the non-advanced track during freshman year of high school, then my parenting has played a role in determining not only my child's courseload, but also the kind of children with whom he or she is interacting. You can distinguish conceptually between parents and peers, but the line blurs when we get into explaning child outcomes.

Comments for this entry have been closed
Return to top