Dysgenics: within and between

And the future of humanity

Lyman Stone has a new post over at Institute for Family Studies: Are We Headed Towards ‘Idiocracy’? A Look at ‘Dysgenic Fertility’. His main point is that the strength of in within group dysgenics is overrated by some. He notes:

The first place to look to see if dysgenic fertility could be an issue is large population genetic registers. In these big datasets of human genetic data, researchers can ask: are genes believed to predict intelligence more common or less common in more recently-born cohorts (younger people) than earlier-born cohorts (older people)? Taken at face value, these studies suggest that the genes associated with higher intelligence seem to be getting less common over time in some countries. Of five studies I could find, studies in Iceland, the U.S.,and the U.K. suggest declines ranging from 0.3 to 0.9 IQ points per generation, while studies of Estonia and Finland suggest no decline and maybe even a slight increase.

In order words, the most direct way to assess the question is to check whether the polygenic scores for intelligence or its proxy educational attainment are actually declining. That is, are the average genetic scores for intelligence (or EA) decreasing between birth cohorts? This question is not so easy to answer because of biases in sampling:

Finally, those studies showing declining genetic scores for intelligence are probably just wrong for a very simple reason. The genes that predict higher cognitive abilities also predict longer lifespans. We also know that the genes that predict higher fertility predict shorter lifespans. The studies I linked to earlier used genetic data collected since the year 2000, and estimate polygenic scores for past generations by taking genetic data from surviving old people.

In other words, even if there is no change between cohorts, the fact that relatively more intelligent people are more likely to survive and end up in the samples of studies mean that it's a foregone conclusion that the older cohorts in a cross-sectional study will have higher genetic scores for intelligence. He appears not to know that researchers were already aware of this bias and corrected for it. Here's Kong et al 2017, the Icelandic study:

For 129,808 genotyped individuals born between 1910 and 1990 POLY_EDU shows a notable and highly significant decline with yob (−0.0182 SU per decade, P = 5.8 × 10⁻³⁵). Average polygenic scores calculated for 10-year bins are displayed in Fig. 2. The relationship between POLY_EDU and yob exhibits nonlinear behavior (i.e., the downward slope seems to be steeper in the earlier years). When a quadratic fit was performed (blue line), the quadratic term of yob is significant (P = 1.7 × 10⁻³). A closer examination suggests that the nonlinear behavior mainly reflects a survival effect rather than a birth cohort effect. The samples studied here were collected between 1998 and 2014, with a majority (68%) ascertained before 2006. For 85,520 of the latter, survival data at 2016 are available. The death rate overall is 19.4% (16,610/85,520) and is 54.5% (13,954/25,610) for those with yob before 1940, compared with 4.4% (2,656/59,910) for those with yob ≥ 1940. After adjustment for sex, yob, and age at ascertainment, each SU of POLY_EDU is estimated to increase the odds of survival by a factor of 1.083 (P = 2.5 × 10⁻¹¹). The positive effect of POLY_EDU on survival is not surprising because it is significantly associated with many other behavioral and health-related traits in Iceland. For example, POLY_EDU is positively correlated with high-density lipoprotein levels, and negatively correlated with triglyceride levels, body mass index, glucose fasting levels, and amount of smoking (P < 1 × 10⁻³⁰ for each of these five quantitative traits; Table S3). Because POLY_EDU has a substantial impact on lifespan, when the samples were ascertained, there would be a positive ascertainment bias, particularly with those born before 1940, for those with high polygenic scores due to the greater likelihood to be alive at the time of ascertainment than those with low polygenic scores. This survival effect has a real impact on the difference in POLY_EDU between the young and the old in the population at any given time. However, for the purpose of estimating the change of the average polygenic score over time with respect to birth cohorts, this can be a source of bias. This bias is expected to be small for individuals with yob ≥1940. Using the latter, the estimated rate of decline of the average polygenic score is −0.0122 SU per decade (P = 2.4 × 10⁻⁷, SE = 0.0024) (red line in Fig. 2). For comparison, we computed two other polygenic scores based on meta-analyses for height and schizophrenia. The polygenic score for height is not significantly associated with yob (P ≥ 0.5). The polygenic score for schizophrenia is estimated to decline at a rate of −0.0078 SU per decade (P = 1.1 × 10⁻³, SE = 0.0024) for individuals with yob ≥1940.

Thus, because they collected their data many years ago and have follow-up data, they were able to estimate the survival effect and thus correct for the survival bias. The red line is the approximately unbiased decline rate for this genetic score. Beauchamp's study of US data likewise noted the issue and conducted subgroup analyses to show that the decline is still present if one focuses on the relatively younger part of the sample.

With regards to the Estonian and Finnish studies, neither of them actually investigated this question and in the Estonian study, it's unclear what evidence in the study he looked at. My guess is that it is this figure:

The figure shows the mean EA genetic scores by place of birth and migration pattern. The results show that people born in the capital are getting genetically smarter over time relatively to the countryside births. This is not surprising, as talent has been moving to cities for centuries. However, one cannot infer the population means from this without weighing these by the changing population sizes. I've heard that someone is working on this Estonian sample to figure out the answer and the answer is that, yes, it shows the decline as well.

The Finnish study has a table showing the genetic score means by birth cohorts. However, there is not enough digits to know whether there is a decline or not. Fortunately, the authors were nice enough to send me the table with more digits. There's 2 samples, and thus 2 tables:

There is a decline over cohorts here, except for the anomalously high 1980s cohort. Second table:

The second sample shows an increase in genetic scores over time. Odd!

Someone with access to the dataset will have to analyze things more carefully to see if this resulted from some other kind of sampling bias. As it is, the larger first sample mostly supports a decline, and the second half-sized sample shows a eugenic pattern.

While we wait for the analyses of the Estonian and Finnish data, we can reassure ourselves that the results cannot really surprise us. We already know from studies of measured fertility and intelligence that intelligence and fertility correlates negatively, which must necessarily lower the average genetic score over time. The point of the genetic score studies is merely to measure how fast this is going in genetic units.

Between groups

Internal dysgenics is quite small, maybe a correlation of -0.10 or so. However, it differs by country. Strangely, Stone did not cite this data, maybe he was unfamiliar with it. Sebastian Jensen conducted a meta-analysis of the association of test scores (regular intelligence tests, or PISA+children proxy method) and estimated the IQ decline by decade:

Here we see that dysgenics is actually stronger in less intelligent countries. Thus, there is a kind of Matthew effect where those who already have less are losing it faster than those with more. And this doesn't take into account brain drain which causes the same effect.

But this is not everything. There is also a very strong negative correlation between the fertility rates of countries and their average intelligence levels. For this reason, at a group level, taking the world population as the population of analysis, there is quite strong dysgenics. Richard Lynn published a paper back in 2008 about this, but others have noted it before I think. Here is the current situation as plotted by me:

The association is extremely strong, -0.72. It's mostly linear with the exception of a few African countries in the top left. Also interesting is that there is heteroscedasticity (p < .0001), meaning that the spread around the regression lines. There is less variance among the 95+ IQ countries than expected. This could be interpreted in a fatalistic way as in, no matter how you end up with high IQ, you end up with predictably low fertility. A kind of modernity doomerism.

This national IQ and fertility pattern has been present for a long time, and thus the world IQ has been declining. Various people have plotted this before, but here's a new plot based on UN data (medium forecast):

Under the assumption of stable national IQs over time (relative to the UK), the world average intelligence declined from 90.5 to 85.8 IQ already and is projected to decline to 80.5 IQ by 2100. Since the UN forecasts are unduly optimistic about rich countries regaining their fertility soon, this forecast may be quite wrong:

Regarding the difference between within and between country dysgenics, or internal and external, I think it was Helmuth Nyborg who first named these two aspects of intelligence dysgenics in his 2012 paper The decay of Western civilization: Double relaxed Darwinian Selection:

This article briefly describes Lynn’s view on what makes modern populations rise and fall. It then provides a demographic analysis of what happens to modern sub-fertile high-IQ Western populations when Internal Relaxation of Darwinian Selection (IRDS) combines with External Relaxation (ERDS, in the form of super-fertile low-IQ non-Western immigration) into Double Relaxation of Darwinian Selection (DRDS). The genotypic IQ decline will ruin the economic and social infrastructure needed for quality education, welfare, democracy and civilization. DRDS is currently unopposed politically, so existing fertility differentials may eventually lead to Western submission or civil resistance.

He was writing about immigration to western countries lowering their average intelligence on top of the dysgenics from their fertility rates. Naturally, he was fired for this (the second time!).

Here it should also be mentioned that using the words dysgenics and eugenics relates to value judgments about what is good and what is not. If someone thinks intelligence is of negative value, the current trend would be eugenic. If they think intelligence is unrelated to human value or worth in any sense, the current trends are neither. The reader is invited to substitute their own value judgments by reversing the wordings as needed. Whatever the case, if we are to be a space-faring species, we have to solve this problem.