Heritability of height vs. weight

By Razib Khan | July 30, 2009 7:10 pm

Megan McArdle has a post, Thining Thin, a follow up to America’s Moral Panic Over Obesity. She says:

1. Obesity is increasing in the population, so it can’t be genetic.
Well, average height is also increasing in the population. Does that mean that you could be as tall as me, if you weren’t too lazy to grow?
Twin studies and adoptive studies show that the overwhelming determinant of your weight is not your willpower; it’s your genes. The heritability of weight is between .75 and .85. The heritability of height is between .9 and .95. And the older you are, the more heritable weight is.

I think the analogy between obesity and height is weakened by likely differences in the effect on the variance of the traits due to environmental changes. First, remember that the term “genetic” is very broad, while the term “heritable” is very specific. Heritability is the proportion of trait variance within the population explainable by variance of genes. The said traits are usually thought of as quantitative traits, like height, weight or IQ, which exhibit a normal distribution. To say that a trait is .95 heritable does not mean that it is caused 95% by genes, that’s not even wrong. Rather, it is to say that 95% of the variance within the population can be accounted for by the variance of genes within the population. But heritable traits are also usually affected by environment; if you starve someone they will be short, but retain five fingers. The number of fingers you have on your hand is not heritable, because there’s no real variance within the population of the trait. It’s genetically specified, but not heritable.

In regards to height and weight, there are some specific details of difference that are important. It does seem that in actuality the average height of American whites has stopped increasing, and are about what they were a century ago (Northern Europeans keep getting taller). Beyond a certain level of nutritional input you simply aren’t going to get any taller. (there are some discussions as to whether high levels of milk consumption are responsible for Dutch height, so what as opposed to how much may matter) Environmental inputs may have reached the era of diminishing returns on height in the developed world. The reason that the heritability of height is lower in poorer nations than in wealthy nations (where it is almost 1) is that in the former there is still environmental factors which account for some variance in height. Some people in less developed nations don’t get enough food to realize their full genetic potential. Naturally the same applies to body weight, heritability should increase as there is an equalization of food intake between the poor and the rich, so that more of the remaining variance will be accounted for by genes because the genetic variance remains the same even when the environmental component is removed from the equation.
Though I suspect we’ve come close to the point of zero diminishing returns on the population wide environmental inputs which might affect height in the developed world, we’re not there when it comes to mean population weight (or BMI). After all, people can take weight gainers or go on diets and exhibit a lot of temporal variation in phenotype in a way that does not occur with height after adulthood (aside from the small but non-trivial gradual decrease in height due to spinal compression). While I assume that the environmental parameter of most importance for height is food, weight (or muscle mass) can be changed through both physical activity and the quantity and nature of food. So if there is such potential plasticity in weight why is it that the heritability is so high? The important point to consider is that if the environmental background shifts the whole distribution can “move.” So societies or populations can get heavier in the aggregate but retain the same heritability. The median weight might vary but the rank order remains the same. The same can be said of height, but it likely there are more “natural” constraints in terms of how tall the average population may get before environmental inputs exhibit diminishing returns (in this way height may be more of a threshold trait where a “good enough” environmental background can result in expression of full genetic potential).
Finally, I suspect that the exact nature of how genes effect height and weight differ as well insofar as I suspect there are more mediating factors in the latter than the former. David Kessler has recently been arguing that modern processed foods create a neurochemical response in the brains of humans which results in “food addiction.” Whatever the truth of this particular claim, Kessler notes that some people seem to be able to control their impulses, or at least do not exhibit a strong fixation on particular tastes. He has suggested that there might be a heritable component here, and I think it gets to the likelihood that a lot of the heritability doesn’t have to do with obesity or weight related genes like FTO, but with variation on cognitive traits which result in a propensity toward certain behaviors which result in weight gain.
So despite the high heritabilities of both height and weight, I do think that the two traits are qualitatively different enough that a tight analogy may mislead. There are still strong class and region based differences in weight, while those for height have mostly disappeared in the United States.
Addendum: I’m not addressing whether or not there is an “obesity epidemic.” I’m not taking any sides as to why people are getting heavier, on average. Also, to be clear, I think a lot more of the range of weight is subject to gene-environment correlations than the range of height. To a far greater extent than height I believe weight has a strong biobehavioral component in terms of its determination.

  • marciano tatti

    don’t want to be annoying but polidactily is inherited

  • http://scienceblogs.com/gnxp razib

    right, but is it heritable? 😉 your point is well taken insofar as these terms are fuzzy on the margins, but it isn’t part of normal human variation. there’s isn’t a normal distribution of finger number. additionally it’s often a single gene pathology. so yeah, it is a genetic trait which is inherited. but it’s not what you normally think as a “heritable” trait.

  • lol

    when heather wrote:
    “overwhelming determinant of your weight is not your willpower; it’s your genes”
    she was likely wrong, but at the very least the data doesn’t say one way or the other. it can be both your willpower and your genes depending on the causality. for example, genes which affect willpower might affect weight. unless someone measures it was can’t say that willpower isn’t *how* genes affect weight.
    what is true is that variation in genes explains most of the variance in weight, whereas the effect of growing up in the same household doesn’t independently make siblings much more similar in weight than their genetic similarity predicts.

  • http://scienceblogs.com/gnxp razib

    i think a *nudge* strategy might be more viable with weight than height. OK, perhaps it would need to be a *push* or *shove* strategy :-)

  • bioIgnoramus

    McGargle’s posts attract numerous commenters with strong views on obesity and its causes, but none who seemed inclined to answer a request for direction to the evidence that was originally used to categorise BMI values into underweight, normal, overweight and obese. Personally, I’d also like to know the evidence that led to “five a day”.

  • http://scienceblogs.com/geneticfuture/ Daniel MacArthur

    Perhaps a *roll* strategy would be most effective?

  • Anthon

    Excellent post. You suggest that the rank order of weight (or stature) is not changing, but that the distribution is shifting. A testable hypothesis might require data on the weights of subjects and their parents/grandparents. Has this been done?
    Similar issues are seen in the secular rise in IQ (the “Flynn effect”): IQ is heritable, but the secular increase is too rapid and widespread to be caused by changes in the the genome. One would suspect that here, too, the rank order is preserved.

  • http://scienceblogs.com/builtonfacts Matt Springer

    “overwhelming determinant of your weight is not your willpower; it’s your genes”
    What, genes don’t affect brains anymore?
    My money on obesity is that obesity is rising because fattening food tastes really good and is now very cheap. I also expect it’ll continue to be a problem until some enterprising scientist invents a little metabolizing machine to sit in our bloodstream and pull out the excess energy. Which might not be that far off for all I know.

  • http://scienceblogs.com/gnxp chet snicker

    Perhaps a *roll* strategy would be most effective?
    is such commentary worthy of one who would deign to impartwisdom upon others? but for the grace of god go i.
    yours truly,
    c.v. snicker

  • http://lowcarbartandscience.blogspot.com agnostic

    Environmental inputs may have reached the era of diminishing returns on height in the developed world.
    That could be — they have to at some point — but showing this requires that the input be monotonically increasing. (“See, we keep boosting X, but Y hardly budges anymore.”) But in most of the developed world, there has been an anti-fats hysteria since the 1970s, and our diet has changed substantially at least since then (in some cases, since the 1940s):
    The foods that make us sturdy — red meat, eggs, and dairy — have been plummeting since 1970, 1950, and 1940, resp. So the diminishing returns idea is premature.

  • http://scienceblogs.com/gnxp razib

    assman, there’s a simple way to test the height issue re: *what* you eat, survey indian americans. there are many families where some siblings remain vegetarian while others have switched whole hog to a high-meat diet.

  • Radar O’Reilly

    Razib, I remember attending a conference about eight years back where a group of scientists that longitudinally studied a cohort of kids from a relatively isolated Central American area. The kids were lean machines by traditional metrics, but the study went on to show that when these kids grew up and wanted a big city life, they adopted big city lifestyles (more meat, sedentary, etc) their cholesterol went through the roof, obesity, and other unpleasantness. The original kids (now adults) remaining in their village remained in excellent health (again, traditional metrics, cholesterol, inflammation markers, BMI/Waist-Hip etc).
    Albeit, it was sponsored by a chocolate company trying to tout protective qualities of cocoa-based antioxidants. But cool idea that was a decent proof-of-concept of these studies. Finding groups like this seems to be the challenge.

  • RT

    “Twin studies and adoptive studies show that the overwhelming determinant of your weight is not your willpower; it’s your genes.”
    McArdle doesn’t think very deeply. Unlike height, which is determined by genes interacting with diet, I suspect that weight (obesity) is largely determined by food consumption which is influenced by willpower which is affected by genes.
    Re the New Yorker article you cite, it is based on the work of John Komlos, a short (5’7″), angry Amercian economics professor at the University of Munich who seems to have a political ax to grind against the US social and economic system. He tends to publish his work in a journal that he himself founded and edits. He’s desparately trying to show American height deficits and argue that this means we need more socialism and a nationalized health system. He dismisses the possibilitiy that genetics could play a role in average height differences between populations.
    If you actually look at the data, male British settlers in 17th C. N. America seem to have averaged around 1.68m. Native-born whites born in the 18th C. averaged about 1.72m and those born in the first half of the 19th C. averaged about 1.73 or 1.74m. As the US industrialized and urbanized in the 2nd half of the 19th C. height hit a low of about 1.7m, and then began to rise in the 20th C. According to NCHS data, men born in 1910 reached an average adult height of 1.73m and average height increased until the 1950 birth cohort, which averaged about 1.78m, and then remained stagnant until the mid-1970s birth cohorts which then saw a small increase to 1.79m.
    Developed European countries don’t seem to be getting taller anymore either, but the trend stopped later in Europe (usually in the late 80s or 90s). Komlos also likes to compare apples to oranges by selecting only the tallest European countries for comparison. He notes that Germans, Czechs, Norwegians (all avg. 1.8m for young men), Swedes, Danes (both avg. 1.81m), and Dutch (1.84m – the only population that is much taller than American whites) are taller than American whites, while forgetting about other countries with socialized medicine that are shorter than American whites such as the Brits, Austrians, Swiss (all 1.78m for young men), French (1.77m), and Italians (1.77m north of Rome, 1.74m south of Rome). Also, Komlos’ theories about height being solely a product of affluence and an egalitarian wealth distribution fail to account for why the Dinaric Alps (Montenegro, Dalmatian Region of Croatia, Herzegovina, and part of Albania), one of the poorest regions in Europe, has an average height of nearly 1.86m for young men.

  • Max

    Genes have a lot to do with it, and so does willpower. But, I can consistantly eat less than my boyfriend, and be bigger than him. I am not overweight, but could be if I didn’t work out lots. Point is, the same willpower (to control amount that and what I eat (leading to the same consumption) leads to a larger weight in myself than my other.
    Ultimately tho, if you eat healthy foods and exercise, you will maintain a healthy weight. It may be harder to do so for some for genetic reasons, but it can be done.

  • Neil

    In the UK, at least, one of the hurdles a study has to jump to access charitable (WT) funding for large-scale genetic association testing, is whether the investigators can claim “their” disease is heritable – usually demonstrated by a high-ish sibling relative/recurrence risk.
    The first WTCCC paper neatly reduces this evidence (with my emaphasis) to:

    The […] conditions selected for study are all common
    familial diseases
    of major public health importance both in the UK and globally, and for which suitable nationally representative sample sets were available.

    which is how I now mentally re-write any claim of heritability.
    WRT the original post, Megan McArdle’s straw man is (as others have suggested) unhelpful. If we changed:

    1. Obesity is increasing in the population, so it can’t be genetic.


    1. Obesity is increasing in the population, so it can’t be entirely genetic.

    would anyone have a problem with it? i.e. it simply becomes a statement that the gene pool is not changing (much), so, without abandoning genetics, we should we should look for additional factors that have lead to the western world’s diabetogenic society.

  • Jazz

    So if the heritability of g is 0.5 then environment plays a significant part?


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar