Skip to content

“The Role of Nature versus Nurture in Wealth and Other Economic Outcomes and Behaviors”

Sandra Black, Paul, Devereux, Petter Lundborg, and Kaveh Majlesi write:

Wealth is highly correlated between parents and their children; however, little is known about the extent to which these relationships are genetic or determined by environmental factors. We use administrative data on the net wealth of a large sample of Swedish adoptees merged with similar information for their biological and adoptive parents. Comparing the relationship between the wealth of adopted and biological parents and that of the adopted child, we find that, even prior to any inheritance, there is a substantial role for environment and a much smaller role for pre-birth factors and we find little evidence that nature/nurture interactions are important. When bequests are taken into account, the role of adoptive parental wealth becomes much stronger. Our findings suggest that wealth transmission is not primarily because children from wealthier families are inherently more talented or more able but that, even in relatively egalitarian Sweden, wealth begets wealth. We further build on the existing literature by providing a more comprehensive view of the role of nature and nurture on intergenerational mobility, looking at a wide range of different outcomes using a common sample and method. We find that environmental influences are relatively more important for wealth-related variables such as savings and investment decisions than for human capital. We conclude by studying consumption as an overall measure of welfare and find that, like wealth, it is more determined by environment than by biology.

Black is speaking at the sociology department this Thursday. The topic seems important.

A few quick thoughts:

1. The article discusses various causes of the intergenerational transmission of education and income, including genetics, direct financial inheritance, and other influences of parents on education, financial opportunities, preferences and attitudes. That’s all fine but I think they missed a few things, including fetal environment (not really the same thing as parents “investing in their children’s human capital”), social environment (children influenced not by parents but by siblings, schoolmates, etc.), and politics (rich people setting up the system in a way that benefits rich kids in general, not just their kids). This last item is particularly salient to me as a political scientist—and also as a statistician, as a reminder that we can consider aggregate treatments when looking at individual outcomes.

Speaking more generally, the division into “biology” (defined as “genetic inheritance”) and “environment” (defined as things that parents do) is problematic.

To give the above list of possible factors is not to say that the conclusions of the above-linked paper are wrong, just that the theoretical framework of the paper could be expanded, which would affect the implications of the empirical results that are presented.

2. I suppose that another key question is how much these results generalize from Sweden to other countries.

3. I didn’t try to follow all the details of the above-linked article. In any revision I would hope to see some graphs of raw data and fitted model. Whenever there is an empirical result, I like to see the trail leading from data and assumptions to substantive conclusions. That should not be difficult to do for this paper, and I think it could make the reasoning clearer and the findings more influential.


  1. Michael Nelson says:

    I have an observation that is a far weaker critique than those Andrew makes, but it still points to further steps that can be taken in future studies to bolster the conclusion that wealth and poverty are not genetically predetermined. The authors’ analysis seems to assume that the distribution of the interaction between wealth and genetics is relatively consistent, continuous and symmetrical, which need not be the case. Suppose, for example, you were to look at only the families in the study’s sample who are in Sweden’s top and bottom wealth percentiles, defined at whatever reasonable level you like but, crucially, defined across the total population of Sweden as a whole, not within the sample. Now put counts into a 2X2 contingency table that compares the binary wealth of a child’s biological parent (top or bottom %) with the binary wealth of the child’s adoptive parent. I’d expect a very small count in the cell for children whose biological parents are very rich and whose adoptive parents are very poor. That is, very few rich families put a child up for adoption who is subsequently adopted by a poor family. Most plausible bivariate distributions of biological wealth and adoptive wealth imply that inferences from the other three cells are sufficient to account for the missing fourth cell.

    An advocate for the genetic propensity for wealth, however, might claim that the effect only goes one way–kids of wealthy biological parents inherent the gene and prosper even in an environment with few resources, but poor children who lack the gene will still prosper as much as their adoptive siblings who have the gene because the environment compensates. Or that there’s some other funky distribution that doesn’t appear unless you have a large number of poor families adopting the kids of wealthy parents. (After all, an imaginary effect can have any distribution you want.) If my expectation about the rarity of this subgroup is accurate, the present study’s design would not rule out that argument.

    Again, this is a weak, and to my mind implausible, objection to the authors’ design and interpretation. But in my experience, people pushing a biased agenda by “scientific” means will gladly stitch together a supportive theory from even the smallest gaps in the evidence base. (For example, creationists have based Intelligent Design on their perception that scientists haven’t yet explained how natural selection could produce every single complex structure that is seemingly only adaptive once fully evolved, while climate deniers embrace every out-of-season blizzard to “disprove” global warming.)

  2. Kyle C says:

    All four authors are economists. I would be very interested to learn whether a biologist thought that this is a good method to distinguish “genetic” outcomes from outcomes “determined by environmental factors.” The philosophical issues raised by the study design seem way beyond the expertise of economists. (To the point in Andrew’s second paragraph under point 1.)

  3. oncodoc says:

    Obviously, wealth has a large social/cultural component that makes genetic analysis largely irrelevant. The most brilliant CEO of a highly profitable Fortune 500 company might not stand out if born into a low-tech agricultural society. The way to do a study isn’t through statistical analysis. Take a thousand new borns and distribute them randomly to people. This would clarify the genetic component. The fact that no one wants to put their kid on the John Rawles roulette wheel means that rich people don’t believe in a genetic explanation for success.
    I did read a thoughtful analysis of this problem once:

  4. For an 80 page paper it sure doesn’t have many graphs, and what graphs it does have are repeats of the same type for the most part.

    I’d like to see spaghetti plotted trajectories through time of wealth for children, colored by adoption status and in several panels based on parent wealth.

  5. Steve says:

    You write, “Speaking more generally, the division into “biology” (defined as “genetic inheritance”) and “environment” (defined as things that parents do) is problematic.” That’s an understatement. The problem I have with all of these nature/nurture, behavioral genomics type studies is that the scientists doing them largely ignore the fact that they are studying two different systems. Wealth or economic outcomes are not the outcomes of a biological system, but a social/economic system. And, genes are not an input into that system. It certainly makes sense to say that certain outputs of the biological system that is my body are inputs into that social system, but that is quite different than saying that I could assume that a correlation between an input in one system and the output in another system could possibly represent a transitive property holding for both systems. This is not a nit picky distinction. No one would think that it makes sense to attempt to determine what caused a certain law to be enacted by investigating how the computer works on which the law was drafted. Why do people think it makes sense to study outcomes of a social system by studying the biology of the agents in that system? Maybe there can be some insight from biology, but that is an extraordinary claim that people seem to think is unproblematic. I propose that we start publishing papers on the quantum mechanical sociology. Afterall, we are all quantum mechanical systems, so just apply QM and you should be able to predict the next election.

  6. jim says:

    This is a great example of something that kind of irks me about a lot of social science research. This paper tells us less than what we could have simply reasoned out long, long ago.

    It’s already clear that, while intelligence may be a heritable characteristic, there’s no genetic law that ensures equal intelligence, talent, or success – or the lack of them – from one generation to the next. But there are legal codes that propagate wealth from one generation to the next. And there is a very-long-established back-scratching social system that propagates wealth. Jeff Bezos’ kids won’t have any trouble getting jobs, talented or otherwise. Who wants to piss off the CEO of Amazon (or any influential wealthy person) by not hiring her/his kid?

    The real wonder is that, despite the many advantages the wealthy have, wealth actually **doesn’t** seem to begat wealth. Large family fortunes mostly decline from one generation to the next. Now and again, some kid picks up where her parents left off and grows the family business.

    If you look at the modern technology tycoons, while most of them were born into a family with solid resources, most of them are, by any measure, intellectually gifted people. And there isn’t a lot of public information about other stupendously wealthy people like Sam Walton, Bloomberg or Buffet, but it’s certainly safe to say that they aren’t stupid.

    So, while intelligence won’t make you a billionaire, you probably won’t get there without a lot of it.

    • jim says:

      So I guess the upshot is that we really don’t need genetic inheritance of intelligence or some special nurturing to propagate whatever wealth is actually propagated in a family lineage. A legal code that passes wealth on and a little social back-scratching is more than enough to account for the wealth propagation that we do see.

Leave a Reply to jim