Rich-poor genetic differences are not just about IQ

My last post mentioned that rich and poor are genetically different. When people contemplate this, their thoughts tend to jump to IQ. They assume that “genetically different” means “genetically predicted to have a higher IQ”. That would make sense, because being smart often gets you a higher income. And indeed, polygenic scores for educational attainment (PSEA) do differ by income, or by other measures of socio-economic status.¹ Here’s a graph from our Nature Human Behavior paper:

Polygenic scores for educational attainment among four groups in the UK

The four bars shows PSEA among four groups: people who were born in UK coalfield areas but left them; people who were born outside them and stayed out; people who were born outside coalfields but moved to them; and people who were born their and stayed there. Coalfield areas are usually poor. The people who were born and stayed there have lower PSEA scores.²

But now here’s a polygenic score for getting heart disease:

Same picture: people in the poorest areas have genes which predict a higher chance of getting heart disease.

Here’s a score which predicts ADHD. Same picture.

In general, for numerous polygenic scores, people in poor areas are worse off — for example, more likely to develop some specific illness. (Geneticists are rightly wary of assigning ethical value to DNA scores, but here “worse off” just refers to the obvious fact that most people would rather not have ADHD or heart disease.) This includes scores for mental illnesses like ADHD and depression, physical illnesses like heart disease, and traits related to physical illness, like body mass index (BMI).

The Parable of the Talents: “For unto every one that hath shall be given, and he shall have abundance….”

If you are relatively rich or high status, it’s probably quite nice to think about these genetic differences in terms of IQ. We feel we deserve our intelligence, or if we don’t deserve it exactly, well, it’s still a deep part of us, isn’t it? Maybe we’re lucky, but our IQ still somehow… reflects well on us. (Perhaps academics are especially likely to feel this way.)

It’s harder to tell yourself that story about heart disease risk. Do you deserve to be less likely to die of a heart attack than someone else? Is that “part of who you are”?

It is often bad when debates about genes get reduced to debates about IQ. In this case, thinking about the genes/socio-economic status gradient in terms of IQ is morally comforting, in a misleading way.

Why tho?

We have an idea why rich people might score higher for IQ: brains help you get rich. But why there should be differences for other, health-related scores?

One possibility is that this is all a mistake. Correlation is not causation, even for DNA. At least some of the correlation between PSEA and education really comes from the environment. In fact, part of the point of our paper was that genetic scores are correlated with environments (geographic areas) which might confound our analyses if they are not accounted for. Could that explain everything above? Do these scores work simply because they are linked to poverty, which affects people’s health?

That’s unlikely. Part of the correlation between PSEA and education is causal, we know that from studies of siblings, who share the same environment.³ And PSEA is especially likely to have environmental confounds, because parents try very hard to educate their children, and it’s probably easier to do that than to make your child healthy. Empirically, other polygenic scores indeed seem to have smaller environmental confounds.

Another possibility is that these health scores affect your income just like PSEA. Maybe bad health makes you poorer, because it’s harder to keep a job up, or because you have to spend money on your health.

A third possibility works via the marriage market. Suppose some of these polygenic scores correlate with attractiveness. For example, tall men are often attractive; so are smart people; maybe polygenic scores for heart disease correlate with visible signs of health, which are attractive. If so then people with high scores on one dimension will tend to partner with those who score highly on another dimension. Tall men marry smart women, say. Then their children will have more of both traits, and so, over generations, all of these attractive genetics get associated.

That story might be testable. It predicts that visible scores will correlate with each other (and with socio-economic status). If there were a polygenic score that didn’t correlate with visible phenotypes — say, it made you more likely to get a disease, but it didn’t make you look physically different — then it wouldn’t correlate with the other scores, because it couldn’t be attractive in marriage markets.

My main point, though, is that genetic differences between rich and poor are multifaceted. They relate not just to intelligence but to many dimensions of health. Remembering these other kinds of difference might give us a bit of humility about the luck of the genetic draw.

---

If you liked this content, then I would love you to do three things:

• Subscribe to this newsletter. It’s free.

• Share Wyclif’s Dust on social media. By telling your friends and/or followers, you’ll be doing me a huge favour.

• Read about the book I’m writing. It’s called Wyclif’s Dust, too. You can download a sample chapter.

1. Educational attainment is not IQ, but it’s close enough. In fact, the polygenic score that predicts educational attainment is better at predicting IQ than the polygenic score that was designed to predict IQ. But if you think of PSEA as predicting non-cognitive skills that help someone do well at school, my argument will still make sense. Relatedly, see this paper and this commentary by Scott Alexander.

2. I could have used individual-level measures of income and they would tell a similar story. I’m using geography just for convenience.

3. And whose genes are randomly assigned by the “lottery of meiosis”.