Genes and health insurance coverage

We provide the first investigation into whether and how much genes explain having health insurance coverage or not and possible mechanisms for genetic variation. Using a twin-design that compares identical and non-identical twins from a national sample of US twins from the National Survey of Midlife Development in the United States, we find that genetic effects explain over 40% of the variation in whether a person has any health coverage versus not, and nearly 50% of the variation in whether individuals younger than 65 have private coverage versus whether they have no coverage at all. Nearly one third of the genetic variation in being uninsured versus having private coverage is explained by employment industry, self-employment status, and income, and together with education, they explain over 40% of the genetic influence. Marital status, number of children, and available measures of health status, risk preferences, and prevention effort do not appear to be important channels for genetic effects. That genes have meaningful effects on the insurance status suggests an important source of heterogeneity in insurance take up.

That is from a paper by George L. Wehby and Dan Shane.  Via the excellent Kevin Lewis.  We do need to know more, but one possibility is that the adverse selection model of health insurance is much overrated, and advantageous selection into health insurance is a live possibility.


We provide an investigation into whether and how much legislation explains health care coverage.

Using a pre and post ACA test, we find that what explains explains health care coverage, or not, and health insurance product variability (guaranteed issue and essential coverage requirements) is the passage of the ACA.


FYI, had two recent posts that were enlightening. First, cut out the middleman and second, post the frickin' prices for the love of all that's holy. I can't think of another industry that does this. They don't do it because they're afraid of a deflationary spiral and for good reason.

Mods. MODS!!!!

if the ACA was good would it need to be mandated???

You understand the economics of pooling risk, adverse selection, and private information, do you?

I understand that the ACA was a tax on poor young people, to be transferred to rich old people.

Young people grow older, in case you didn't know, so the benefit is shared. If you are 26+ now, you will be 35 in later. That's the concept of a large bracketed pool. You will not get your social security retirement benefit when you are 25 years old either.

Grow up. You will grow older with time. Hopefully, wiser, too.

If you want to learn about social insurance, you might want to go to Robert Shiller's Economics and Finance course at Yale. Go to where he has a discussion of this.

Also, Anon, if you have a pre-existing condition at 26+ you will not be denied coverage. If you understood adverse selection and pooling, you would also understand a bit more about mandated coverage (or penalty for non-coverage). Either you know about this and are just trolling, or you are ignorant.

Yeah the 2nd "Anon" was not me. My point is "pooling" is great all insurance companies do it. But why mandate it??? Why not offer a health insurance program that is fiscally responsible, provides a decent level of health care and is OPTIONAL. In other words make it good and people will want it. The problem is that our politicians, at least those who are designing our health care solutions, are socialists. Because of this their efforts are doomed to fail and we are doomed to endure a failing system until we revolt against it. Why not make it right the first time???

If indoor plumbing was good would we need to mandate it?

Pre-ACA more than 80% of the population under 65 had health insurance. We find that ACA accounts accounts for only a relatively small percentage of health insurance coverage despite such common insurance issues.

Only in the US can such a study be even conceived...

Behavior genetics deals with the causes of variation of individual traits, attributes etc. in a population. If there is no variation in the trait, there is nothing to be estimated. If you tried to run a similar twin model as in that study in some country with universal health care coverage, the model would not converge because the variance of the dependent variable would be zero.

this idea of health insurance as insurance is pretty uniquely American too

Health insurance is common in European and Asian countries. It is either used to fund universal systems or to pay for private health care that supplements public health care.

What is interesting is that there is no behavioral genetic explanation which shows why Canada, the UK, Denmark, and others provide universal health care to citizens.

Interestingly, there's also no astrophysical explanation of that. It almost seems as if a given science cannot provide answers outside of its domain.

I have bad news for you: in countries with national healthcare systems, things like obesity, longevity, health outcomes, healthcare utilization, exercise, and the lot are still highly heritable. SES and IQ correlations are also large, and things like the Swedish lottery natural experiment show no effect of monetary transfers on health. As Gottfredson notes:

"Great Britain and other countries that had expected to break the link between class and health by providing universal health care were dismayed when the disparities in health not only failed to shrink but even grew (see The Black Report by Townsend & Davidson, 1982; also Link & Phelan, 1995, p. 86; Marmot, Kogevinas, & Elston, 1987, p. 132; Susser, Watson, & Hopper, 1985, p. 237)."

Ok, so if genes can explain 40% of the variation in health insurance coverage, then what shall we make of the apparent percentage of intelligence controlled by genes? It's not like there's a gene for health insurance, specifically. Rather there is a gene for some sort of personality factor that increases the likelihood of being able to buy health insurance and/or wanting to.

Similarly, much of the variation in intelligence attributed to genes may not be directly attributed to "genes for intelligence". Rather the genes may code for some other attribute that influences the development of high intelligence. For instance, maybe skinny guys aren't good at sports leading to a greater focus on scholarly studies. Or cute (facially-symmetric) children receive more attention from adults leading to more cognitive development.

IQ is stable by age 6. So there goes your first theory. Unless toddlers are on gear for some reason.

To the second one, that causal mechanism makes little sense. The correlation is there, but it’s long established science that it’s an outcome of genetic load causing both more facial asymmetry and lower IQ.

Those were just random examples not intended to be used as actual theories. The point is that "attributable to genes" doesn't necessarily mean directly attributable. There could be some very circuitous route by which genes impact the development of intelligence, and that route may involve environmental feedback.

Sure. Exposure to lead is one theory.

Unfortunately we’ll soon find out the validity due to gross governmental incompetence in Flint. There’s a PhD thesis there waiting to be written in another 20 years.

Outside of deformity, malnutrition, head trauma, and poisoning though, it’s safe to say we have enough twin adoption studies to say, conclusively, no.

Heritability with zero shared environment is now at over 86%. That percentage goes up every few years with higher quality data and studies.

It’s not roundabout. It’s not circuitous. And it’s clustered (not bivariate) with facial symmetry, height, and athleticism. The memes with smart, short anti-social smart nerds were wrong. The truth is closer to smart, tall, athletic, high IQ and high earning. The outside Venn diagram of high IQ and nothing else inhabit the world of academia. A weirder combination of above average IQ and nothing else tend towards government and journalism.

Again, the "heritability with zero shared environment" twin studies are based on the same methodology that leads to results like 40% genetic causes for health insurance coverage. There might be an underlying gene that influences it, but it could be gene->behavior->environment->intelligence, not gene->intelligence.

That just further illustrates my point - finding that 40% of health insurance coverage is attributable to genes casts significant doubt on the validity of the methodolgy of these studies. In general, not just this one.

Hormones. I strongly believe that hormones are one of the most underrated phenomena on Earth. So much of our behavior is dictated by our individually esoteric mix of baseline and environmentally modulated dopamine, serotonin, oxytocin, etc... Non-identical twins have a different mix of hormones and receptors. That’s a big part of why they have different risk-aversion profiles, and why they show significant infra-familial variance in risk-aversion. I also believe that hormones explain a lot of bad parenting, differential motivation, differential emotivity, loneliness, etc...

I believe a lot of the biggest unsolved human-centered problems come down to hormonal variability. Once we get cheap, portable brain scanners and miniaturized omni-hormonal detectors to study differences in all sorts of behavior, I believe my theory will be validated. What’s the prescription? Well, changing the environment can only do so much. Prescribing a change in behavior assumes that the person has the right mix of hormones + hormonal receptors to follow your prescription. CBT is great but it can only go so far. I think the next leap is endocrine programming, like using CRISPR and/or technologically assisted pharmacological interventions (implant that reads and dispenses multiple types of hormones in small quantities). That leaves the issue of: what’s the ideal mix of hormones? And is there an individually optimal mix of hormones that has negative externalities (e.g. if all men had high testosterone)? Are there agency issues in trying to control for those negative externalities, and would endocrinal modulation be a societal net positive even if we let people choose their hormonal profile (I suspect so)?


Yeah, every behavioral trait ever studied has shown some genetic influence. Literally every trait studied. And genes typically explain about 50% of the variance. So add this one to the file.

It's strange, economists are moving (running?) away from economics into the hard sciences, while leaving behind a wasteland of economic theories that don't reflect reality. Let's hope that economists don't do to the hard sciences what they did for economics. Is that too critical an assessment of the current state of economics?

One of my all-time favorite articles is David Hemenway's QJE article about propitious selection, cited in the advantageous selection link. They say that recent literature has favored calling it "advantageous selection" which seems like a more boring phrase as well as one that fails to properly cite Hemenway's work.

This melding of Neoliberal politics with obsessive eugenics is truly terrifying. I can't even imagine the horrors it will unleash.

To what extent is "genetics" here a cover for "people from foreign countries not legally in the country and able to obtain health insurance through legitimate channels"?

Can't Kevin Lewis become something besides excellent?

Comments for this post are closed