Garett Jones’ paper

Plamus linked to it in the comments section and I’ve seen it linked elsewhere as well, it’s an interesting paper.

Here’s the abstract:

“A recent line of research demonstrates that cognitive skills—IQ scores, math skills, and the like — have only a modest influence on individual wages, but are strongly correlated with national outcomes. Is this largely due to human capital spillovers? This paper argues that the answer is yes. It presents four different channels through which intelligence may matter more for nations than for individuals: 1. Intelligence is associated with patience and hence higher savings rates; 2. Intelligence causes cooperation; 3. Higher group intelligence opens the door to using fragile, high-value production technologies, and 4. Intelligence is associated with supporting market-oriented policies. Abundant evidence from across the ADB region demonstrating that environmental improvements can raise cognitive skills is reviewed.”

I don’t buy 4 at all unless/before much more work is done in that field. Now it mostly just reads ‘I read Caplan’s book and people I know talk about it so I should probably mention it in my study’ to me. The other parts I don’t have strong opinions about. Below’s some stuff from the study and my remarks. Here’s Figure 1 from the paper, you have log-GDP pr. capita up the y-axis:

The ‘PRC’ in the corner is China, and there are plenty of reasons (the name of the most significant one is Mao) why you’d think it makes good sense that they haven’t managed as well as the theory suggests. The IQ-effect is huge: “Jones and Schneider […] found that across countries […]: 15 IQ points is associated with a 150 percent increase in productivity.” If you think simply in terms of labour input, this finding would suggest that in a country with an average IQ of 115, 2 average workers can be expected to add the same value to a product as (‘do the work of’) 5 workers living in a country with an average IQ of 100. Yet the private returns related to that productivity difference is very small; in the paper they mention an estimated wage differential of just 13 percent.

There’s a lot of stuff in the paper, I’ll just go through a few interesting bits I found. Here’s some stuff on environmental factors and their influence on IQ:

“there is a vast public health literature on environmental correlates of intelligence, and many of these papers study nations in Asia. A study of excessive fluoride in Indian drinking water found a 13 IQ point-difference between children “residing in two [separate] village areas of India with similar educational and socioeconomic conditions” (Trivedi et al. 2007, 178). If even half of this relationship is genuinely causal, and if intelligence has some of the technological and political spillover effects discussed below, then public health matters are of first-order concern for economic development.”

The impact of just two environmental factors of that size could in theory reduce the mean intelligence of a population with Mensa-level average IQ to that of current-day Japan. These effect sizes are huge.

“Arsenic and fluoride exposures are also associated with low IQ in the People’s Republic of China’s (PRC) Shanxi province (Wang et al. 2007, 664), even when comparing “groups [who] lived in rural areas with similar geographic and cultural conditions and a comparable level of socioeconomic development.” High arsenic exposure was associated with a 10-point IQ gap, and high fluoride exposure with a 4-point gap. In both cases, the “normal” group had an IQ of 105, 5 points above the US mean.

In the Visayas region of the Philippines, Solon et al. (2008) found evidence that lead levels reduced the IQ of children. In their study, one microgram of lead per liter of blood was associated with a 2.5 point reduction in the verbal IQ of older children, and a 3.3 point reduction in the IQ of young children. In their sample of children, the levels of lead in the blood averaged 7.1 micrograms per liter, so lead exposure could be costing the average child in this sample 15 IQ points even under conservative estimates.”

The role of nutrition is mentioned in the paper, but they don’t go much into the specifics. I’m pretty sure that’s one of the main things holding India back on the IQ-scale of Figure 1.

I think both point V and VI are only/mainly there because of the agenda of the authors and I hate that kind of thing. V is almost pure speculation using an already (with respect to which conclusions can be drawn from the findings) speculative voter preferences model from the US to talk about East Asia. Smarter people will be more likely to support free market policies if they think they’ll gain from it and they get a say in the matter, which depends mainly on how the local government decides to split up the cake. Show me a group of American professors of theoretical physics pushing for more free market policies in education (fewer gov. subsidies). No, that’s not the relevant margin, but to take an extreme example in the opposite direction, in a standard median voter model you could have an IQ increase of 30 points of the 4 top deciles having no effect on policies whatsoever, if the intelligence of the median voter is unchanged. Yeah, you might argue the IQ effects are to be had on the other side of the distribution, but model symmetry means that you could make the same argument and apply the change to the 4 lowest IQ deciles. Conceptually they probably just take up this subject to encourage further research, but I’m one of those people thinking that Caplan is drawing way too strong conclusions from his findings already, and using IQ proxies to speculate about effects in countries looking nothing like the US, having wastly different political systems – well, that’s just not very smart. Point VI is of the same kind – it smells of ‘we want to push this idea, how can we include it in the paper’-motivation. It mentions one way to increase a country’s IQ – immigration. From the paper:

“Even if scientists and public health officials quickly reach their limits in raising a person’s IQ—again, not a foregone conclusion — we still have a reliable tool for raising a nation’s IQ. Encourage immigration by individuals with higher average intelligence. Many countries implicitly do this by permitting high-skilled immigrants to enter and work legally.”

Nowhere in the paper is it mentioned that this is most likely a zero-sum game. One country’s gain is another country’s loss. And the ‘many countries implicitly do this…’ part is correct but only half of the story, as many countries, especially Western countries, also implicitly do the opposite – import massive amounts of low-IQ immigrants (and also implicitly form/maintain policies which encourage these people to have a lot of children, lowering national IQ and future human capital even further).

July 1, 2011 - Posted by | immigration, IQ, Studies


  1. As you note, there is much in the paper, and a fair bit of it is speculation – thinly proven, and best used as “something to think about”.

    I must, however, take some exception to “Nowhere in the paper is it mentioned that this is most likely a zero-sum game. One country’s gain is another country’s loss.” I would think that only the immediate, first order effect is zero-sum. The children of high-IQ individuals, who are more likely to be high-IQ, should have a better chance of realizing the environmental portion of their IQ-capacity in a richer country, if only for escaping the possibility of malnutrition. Basically, in our Egypt-vs-Switzerland example, a high-IQ couple moving from Egypt to Switzerland is indeed a gain of 2 high-IQ individuals for the Swiss, and a loss of 2 for the Egyptians. But the couple’s child has an higher average expected IQ after the move, and this delta-IQ is a net gain in aggregate.

    Comment by Plamus | July 3, 2011 | Reply

  2. If low average IQ leads to bad institutions, the move has a negative second order effect on the sender country as well – the people living in Squalid Poor Country X will have lower expected future offspring IQ after the move. And they need another genious just to get back to square one. One way to model it would be to say that when smart people move away, the risk of malnutrition ect. of the remaining population increases. One of the main points of the paper is that the societal effects related to IQ are potentially quite large. So let’s say the smart guy in the village pushes for a public well in the home country – 15 children in the village don’t get cholera while young and average IQ of that group increases by 5 points. Or it could work through a dozen other mechanisms (also non-political effects: The move impacts the savings rate negatively, maybe there’ll be more conflict due to the move, better harvest methods foregone, the smart one was a schoolteacher and now the school has to close, ect.). The effect is probably not that large, but how large is it?

    The above is also related to one of the main problems, which is that I don’t necessarily think IQ should be the (only) target variable. Why not optimize the societal externality instead? Even if the dynamic case for pure IQ were clear-cut, which I don’t think it is, why not ask ourselves if it’s IQ we want to maximize over when deciding on immigration levels? Maybe the size of the contribution to the positive population IQ externality of individual i is increasing in the difference between the IQ of individual i and the population IQ – if so, the environmental IQ costs of having individual i stay in SPCX might be more than offset by the higher societal returns to IQ pr. IQ-point in SPCX. Maybe the smart thing to do in the aggregate is actually to have smart people from rich countries move to poor countries?

    Another thing is that the people who stay behind sometimes (/often? /mostly?) have more children than people who move. It’s an open question if the high IQ individuals who move have fewer children than they’d have had if they’d stayed in their home country. For some country pairings this is definitely the case. This could potentially undermine the IQ argument. Three children with an IQ of 120 vs. two children with an IQ of 130 – which outcome do we prefer?

    The static case is zero sum, and even if you include dynamics I still don’t think the net effect is clear at all. Without data this is way too complicated an issue to draw any firm conclusions.

    Comment by US | July 4, 2011 | Reply

  3. These are great points, it’s a pleasure discussing this with you, Sir. If you do not mind, I’ll quote some of them and add my thoughts – not to fisk you statements, but to keep mine somewhat structured and flowing.

    If low average IQ leads to bad institutions, the move has a negative second order effect on the sender country as well – the people living in Squalid Poor Country X will have lower expected future offspring IQ after the move.

    Certainly true. I was focusing not so much on the authors’ proposition – which is intellectually appealing, but far from confirmed – but rather on the traditional rule of thumb that about 30-50% of IQ is inherited, and 50-70% is environmentally determined.

    However, even in the context of the authors’ theory, the effects of a move are ambiguous. You note correctly the effect propagated using the author’s mechanism #1 – higher savings rate. But this effect seems to be the only one of the 4 (maybe #4 – support for market-oriented policies – partially) that does not require a critical mass of high-IQ people. Mechanism #2 – cooperation – certainly seems to be one where a clustering seems to be critical. A genius in SPCX might contribute a lot in disease prevention, agriculture, politics, etc.- but history shows that more often than not such people are ostracized and persecuted if they try to upset the dysfunctional power (political, religious, tradition, etc) status quo. History just selectively reports on the few that succeed, and not the many that fail and are smothered by the dark mobs. Mechanism #3 seems to be even more cluster-dependent. In fact, I’d expect that it is conditional not so much on average IQ, but more on the presence of a certain number of individuals above a certain threshold IQ in an economy. Thus, the mechanism may have a threshold below which it’s mostly inactive – and thus SPCX has little use for a high-IQ individual in foreseeable future, unless they decide to clone that person, or use them for “farming” his high-IQ progeny. Hence to your question:

    Three children with an IQ of 120 vs. two children with an IQ of 130 – which outcome do we prefer?

    I would probably answer option 1 in SPCX, and either in a richer country, but the rich country could better use option 2.

    Maybe the smart thing to do in the aggregate is actually to have smart people from rich countries move to poor countries?

    In an ideal world, maybe. But it appears that low average national IQ unfortunately highly correlates (is it causal? Is this mechanisms #2, #3, and #4, in some hybrid form, at work?) with insular thinking, low tolerance for change, and high distrust for outsiders. The majority (not half, as we established) people with IQ<100 seem to be way happier to be demagogued by a snake-oil salesman, as long as he is one of their own, then listen to, let along accept as their own people from outside their group. For examples, see all the scorn poured on the IMF in places like Greece. People actually buy hook, line and sinker arguments like "Whereever the IMF goes, it's a disaster." – the same logic which should, logically extended, have you keep doctors away from epidemic outbreak centers. Thus, until humanity has blended enough for races, nationalities, religions, etc, not to matter, I do not see artificially relocating IQ clusters as effective.

    To recap (and do some stream-of-conscience self-therapy): your points are very good, but I still believe letting high-IQ people migrate to where they are best compensated (not just financially, but also utility-wise) is the best – certainly for those individuals, but also for the world in aggregate. Social engineering very rarely works (see China), and freedom breeds prosperity. The problems of SPCX's are numerous, but "brain drain" is one the other problems cause, not one that causes their problems. Incentives matter. When SPCX's create conditions that are conducive to prosperity, then will their average IQ climb, not the other way around. If you put a diamond in fluoroantimonic acid, you are not going to get more diamonds – you will just lose the diamond. The acid needs to be diluted for the diamond to even survive, and the best way to do that is not to add diamonds until that level of dilution is reached. To continue this analogy, the best way is not to add water either (big kaboom! ensues). Let SPCX's find their own liquefied sulfur dioxide.

    Comment by Plamus | July 4, 2011 | Reply

  4. Thank you for the kind words. I’ve read most of the day (another ~two chapters in Fundamentals of Genetics…), so I’m a little tired in my head, but I thought a response was in order. A few remarks:

    “traditional rule of thumb that about 30-50% of IQ is inherited, and 50-70% is environmentally determined.” I haven’t heard about that rule of thumb, but my understanding is that heritability is a bit higher than that. Then again it depends on who you’re looking at of course, in general the coefficient of heritability increases with age. A couple of links:

    My motivation for asking some of the questions above was only partly to get at the answers, it was as much to illustrate how much stuff Jones disregards in his paper/leaves out. That’s what makes me angry, you (that is, Jones, not ‘you’ you…) don’t write a paper like that, you don’t just, to take but one example, leave out an obvious negative second order effect of interest and if you find it necessary to argue the dynamic case you can’t just leave out the implicit statics.

    I think that if you agree that IQ is important when it comes to the institutional setting, then you need to allow for the fact that causality goes both ways – if IQ matters for institutions then brain drain is necessarily both a symptom and a cause. If better institutions require higher average IQ to get implemented, then SPCX’s can’t be expected to ‘create conditions that are conducive to prosperity’ if all the smart people leave. Maybe they couldn’t create those conditions anyway – what do I know? ( not all problems have a solution ) but that’s a different point. This would probably be a good time to state clearly that I really don’t have a dog in the fight as I have no influence on immigration policy anyway – and arguing about ‘optimal policies’ on a meta level is not something I much like doing in the first place.

    Comment by US | July 5, 2011 | Reply

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: