Sunday, April 16, 2006

The meek (minded) shall inherit the earth

The disovery of the connection between IQ and per capita income by country was ground-breaking. But there is another correlate of national IQ, and it's even stronger than per capita income: The correlation between IQ and number of births per woman. The correlation is, not surprisingly, inverse. Running the numbers yields an r of .81 (r-squared of .65--see chart below).

I haven't read Vanhanen's and Lynn's tome, but I understand that the r for IQ and PPP was found to be .73. Taking the same IQ scores and running them against the most recent PPP estimates yields an r of .60 (r-squared of .36)--moderated, perhaps due to the recession a few years back and the abrupt climb in oil prices that have bumped up the PPP in the moderately low IQ Middle East, but still significant. So IQ "explains" about one-third of national per capita income and two-thirds of fecundity. One notable difference in the correlations between IQ and PPP versus IQ and births is that East Asia doesn't conspicuously outly in the latter like it does in the former. Babes from Hong Kong average less than a single birth (.95 per woman)--and the place is also the IQ apex (107).

We are taking a deleterious dive. Liberalized economies open up opportunities for careers, travelling, education, birth control, economic independence for women, etc. Making babies has always been an option, but now more than ever there's more desirable alternatives for sharper people. Wealth is tied to IQ, but the correlation between PPP and fecundity is considerably weaker (r of .55, r-squared .30) than it is between IQ and fecundity. The driver appears to be IQ. Women in countries with abundant natural resources but moderate IQs (Middle East most saliently) do not have as much cerebration to keep them from having babies. Religiosity, independent of IQ, is probably correlated with having kids as well.

What to do to counter this trend?

-Higher IQ countries attract migrants from places less cognitively endowed. The correlation coefficient for net migration rate and IQ is a positive .40 (r-squared .16) if countries with a rate listed as zero are excluded (this is only a rough approximation. I cannot find much data on migration rates. Searching online I've only been able to find data parroting what's on the CIA's site, which lists lots of places as having no net migration/emigration even though they clearly do (like Zimbabwe for example)). Stopping the flow will lower the cost of living, bump up wages, and make public schooling more attractive.

-Invest in research to extend productivity as people age. Median age and IQ correlates positively at a very high .85. That is, 72% of the median age in a country can be explained by looking at that country's average IQ. Put in another way, every 1.17 point increase in IQ ups the median age by a year. The planet's up-and-comers have lower IQs. The older folks have the brains. Okay, that's pretty crass but generally true. At least when nation to nation comparisons are made. The Flynn effect is (hopefully) counterbalancing this to some degree, but it may be abating.

-Create incentives for people of moderate affluence to have more children. France has taken a few baby steps in this direction. The US tax code, unfortunately, punishes the fertile wealthy by phasing out exemptions and dependent credits and by the existence of the AMT.

-Spend aid money on birth control in the third world. A family member of mine worked in the diplomatic core of the State Department for several years. In Bangladesh he was involved in a program that distributed birth control devices. The program targeted the most impoverished areas. Critics lambast this sort of strategy as for being eugenic (that's inherently an evil thing for those of you who do not see why this counts as criticism!). But it helps realize a putative goal of the left: Closing the wealth gap.

-Watch Idiocracy. At least you'll be able to approach the future with tongue-in-cheek!

++Addition++John Bolton makes an excellent observation. Fecundity might be a better explanation for disparities in PPP than IQ is. IQ and births are decent proxies for one another. So which one is effecting PPP?

IQ is the primary causal factor. Its correlation with PPP withstands controlling for births per woman, but births per woman loses statistical significance as correlated with PPP when IQ is controlled for.

Here's how it breaks down: When IQ and PPP are correlated (without births controlled for), there is statistical significance factor of less than 0.001% with a correlation coefficient of .60. Each IQ point corresponds with a $653 boost in PPP.

When births and PPP are correlated (without IQ controlled for), the relationship is over 99.99% assured with a correlation coefficient of .55. Every baby costs $5,052 in PPP.

The slightly attenuated correlation coefficient for births versus IQ suggests the latter is the more powerful "determinate" of PPP. But we need to control for each of the variables. When births are controlled for, IQ remains statistically significant at about a 99.5% confidence level. Each IQ point now means a $485 boost in PPP.

When IQ is controlled for, births fall outside even an 80% confidence level (P-value of .78--meaning there's a 22% chance that the relationship is random. (The P-value is basically the same as the statistical significance factor, except the former relates to a multiple regression). In this less reliable relationship, each baby costs $1,752 in PPP.

So births per woman is a close proxy for IQ, and IQ is correlated with PPP. The corollary is, then, that births are correlated with PPP as well, although the correlation coefficient is not as strong as it is with IQ (.60 vs .55). When we control for births in the equation, IQ weakens a little but retains explanatory power (95% CI is the gold standard). However, when we control for IQ, births become only a third as powerful, and unreliable at that.

A rough reverse-syllogism: A couple comprised of two 130 IQers can afford to have a few children and still be considerably wealthier than a childless couple comprised of two 70 IQers. Having children does lower the economic power per family member, but IQ more than compensates for this cost. The dependents are net liabilities until they're no longer striplings. Lynn and Vanhanen should have (did?) considered this.

The most interesting (and tragic) aspect of the analysis is how few children the sharp nations are having (recall the correlation coefficient of .81). While IQ and PPP got the headline, the relationship between IQ and births is less tenuous--it appears to be rock-solid (I'm going to test it with some controls in near future).

Also, I humbly offer the data I'm working with to anyone who is interested (Email me if you are). It is in excel. More alacritous minds may find something my pile of mush is missing.

(Human biodiversity)


faq said...

I'd guess countries with low IQs have shorter life expectancies. The average age likely paints a gloomier picture than necessary.

jeffreyok said...

The purchasing parity estimates are from different years. That could be throwing it off?

JSBolton said...

Here is a question that should be cleared up: how may one exclude fecundity as the prime causal factor, and substitute IQ? Not that it can't easily be done; but, at the same time, population growth rate, fertility and labor force growth rates, can be seen as increasing the difficulty of generating per capita production growth.

crush41 said...


Right you are. In fact, life expectancy correlates just as strong with IQ in a positive direction as the median age does in a negative direction.

When adjusting to "average years left to live" the correlation with IQ stays within a 95% CI (about 97%) but it has a modest r-square of .10. The correlation is actually positively (meaning the average joe in a country with higher IQ has slightly longer to live than the jim in a duller country).

Whether or not this paints a gloomier picture or not depends on how developed countries are going to be able to chug along with a population that keeps getting older. In pure economic terms, the sixth of the EU population that is over 65 (a greater percentage than that of those under the age of 14) is probably better off gone.

Old folks to take care of will slow growth, making it harder for the young in the population to afford children. The correlation between median age and births per woman is an astounding .88 (r-squared .77). My first reaction is to attribute it to the fact that old people don't have kids. But this is not a birthrate--it's the number of children each woman has ever had. If anything, an older population should inflate the number of children per woman but it does the opposite. Causation is tough to pinpoint, but it is sensible to assume that as countries get older, the fertile portion of the population will be less likely to have as many children as they would if there were less elders to take care of.

crush41 said...


I don't think so. Those used in IQ and the Wealth of Nations all have estimates at least as recent as 2004 (from a cursory looking at them) and almost all are from 2005. The older dates are from places like Seychelles.

crush41 said...

John Bolton,

Wow. That is perspicacious and something that has to be considered. I'm smacking myself on the forehead for not addressing it. That addressing is now in the original post.

JSBolton said...

This is probably relevant to what you're doing. ...searhed: lynn vanhanen fertility IQ per capita income...[PDF] The Impact of National IQ on Income and Growth
File Format: PDF/Adobe Acrobat - View as HTML
IQ account for the substantial variation in national per capita income and ... development (Lynn/Vanhanen 2002:188). 12 . Finally the fertility rate is used ... - Similar pages The author uses data disaggregated for male and female education between countries, another source of high correlation to fertility and IQ, or confounder

crush41 said...

John Bolton,

Thanks for that. There's a lot to digest there. I am going to run the Economic Freedom and Democratization numbers. What stands out initially is that there's not much mention of variables after control (I've only perused it but will make it a weekend project). The author contends that because EF and Democratization Index together explain a high amount of PPP variation does not get at the underlying cause. Ditto education rate increases.

JSBolton said...

The additional years of education, come well after IQ has stabilized, so that can't be the cause. Likewise, there are countries which are restrictive of access to higher education, yet this doesn't seem to affect their average IQ, eg: China, NKorea, HK formerly. European countries not long ago, used to send far smaller percentages to college or even non-vocational HS than the US, yet their average IQ's stayed higher.

Anonymous said...

Emmanual Todd thinks there's a three-step:

1) Mass Literacy
2) Revolution
3) Lower Birthrates

Is he wrong? If not, how's the three-step fit?


Audacious Epigone said...

There already are high levels of literacy in many sub-Saharan African countries that have TFRs of 4+ and there are countries like India where literacy rates are barely 50% that do not have high TFRs, so I'm not convinced that the narrative absolutely fits, but that generally seems to be the pattern. The question, of course, is whether or not the subsequent steps follow the one before them only because that step occured, or if something else underlies each of the steps eventually occuring.