The Origins of Savings Behaviour

Bryan Caplan points out a paper by  Henrik Cronqvist and Stephan Siegel on the genetic and parental influences on savings behaviour. The first part of the abstract reads:

Analyzing identical and fraternal twins matched with data on their savings propensities, we find that genetic variation explains about 33 percent of the variation in savings behavior across individuals. Parenting effects on savings behavior are strong for those in their twenties but decay to zero by middle age, i.e., parents do not have a lifelong non-genetic impact on their children’s savings. The family environment when growing up and an individual’s socioeconomic status later in life moderate genetic effects, so that more supportive environments result in a stronger genetic expression of savings behavior.

There are no surprises in these results. The proportion of the variance in savings explained by genetic variation is typical of that for many other social traits. And once opportunity is equalised, genetics becomes more important.

The evidence that parental influence fades out for older subjects and disappears by age 45, compared to the relatively constant genetic effects, is interesting. The break down of effects by age is not a regular feature of studies such as these (it comes at the cost of sample size). The authors write:

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

It is a pity that savings behaviour is only exhibited without significant constraints in adulthood, as if we could also get data for the first twenty years of life, we might also see the genetic influence increase as happens for IQ.

In the conclusion, the authors link the observed behaviour to time preference and self-control. Those with lower savings rates are probably also suffering the consequences of impatience and poor self-control in many aspects of their lives.

One explanation for why savings behavior is genetic appears to be that an individual’s time preferences are partly genetic. Our evidence of a significant positive genetic correlation between an individual’s savings and income growth supports such an explanation. Some individuals are born to be more patient, and this affects these individuals’ savings behavior, as well as other outcomes, e.g., the choice of income process. Moreover, the negative and significant genetic correlation between savings rate and both smoking and body weight suggests that behavioral factors such as lack of self-control may also affects savings behavior. For example, to the extent that a high BMI and obesity may be interpreted as an expression of lack of self-control, we conclude that lack of savings correlates with lack of self-control, and this correlation is mainly found to be genetic.

For those looking for more on the subject, there is a lecture by Cronvqist on this paper on Vimeo:

[vimeo http://vimeo.com/16490797]

 

As an end note, in the closing of Caplan’s post, he notes that there is still much variation that is not explained by either parental effects or genes. Caplan falls back on his old free will argument, but there is a lot to be said for noise. This is a subject I hope to come back to in the near future.

4 comments

  1. Looking at this paper, do economists hate graphs or something? What’s with the reliance on data tables, ew.

    As you point out, this probably isn’t terribly interesting since variance explained is about what you’d expect to see from other related studies (though I agree the decay is an interesting result). The interesting thing to me would be to look at what the important environmental factors are, what the genes are (I assume substantially DRD_), how they’re distributed across the population, etc

      1. Sure, I’d agree that single genes are going to be hard (and probably fairly meaningless because of how they’ll interact…?). But in the GREML paper, I’m surprised they didn’t try to reduce the dimensionality via PCA/ICA/etc. Then the first couple of components should say which groups of SNPs are important, right? I don’t know, is this too noisy in population data to do effectively? I think I may be too used to working on the single animal level…

Comments welcome

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s