While this correlation is pretty intuitive I think something like "standard of living" would be more descriptive rather than "GDP growth". GDP isn't the best metric for the economy because it doesn't distinguish between financial assets and real earned wealth.
The US has been "growing" since the 80s if you focus on GDP only. However the real economy sans financial assets has been stagnant implying the economy hasn't really grown. I think high national IQ represents worker capital like you said and therfore potential for a high standard of living.
IQ is not intelligence. Defining intelligence by IQ gives circular reasoning. b.-Not all countries have IQ records, especially in Africa, Asia, South America. And they are not massive, on the CIA map they took the IQ of 250 people for a country of 190 million people. It is not significant. And it's not the "national" IQ. That means that study is fundamentally flawed and the data is unreliable.
c- If it were for intelligence, the Mayans, Aztecs and Olmecs had an advanced civilization and the pre-Incas when in Europe were in the stone age. The Nubians who were black had a similar civilization to the Egyptians 5000 years ago. Civilization was born in the tropics, not in the north by brown people. Not white. That contradicts this study, since we all have the same intelligence. If not, why didn't civilization begin in Europe?
Excellent post. I read Garett Jones' book "Hive Mind". It is a good book. There are people like Jeffery Sachs, Francis Fukuyama, and James Robinson (of "why nations fail" fame) that study why some countries do better than other. But of course they shy away from topics like IQ. Instead they use euphemisms like "culture" and education in a lot of their work.
I would be interested to read your policy recommendations based on these findings. The reason this type of research is taboo is obviously that people find it offensive and imagine that the policy implications can only be terrible, such as mass forced sterilization programs. The first part may not be changeable, but the second possibly could be. If there were less awful policy implications, people might be more willing to tolerate the offensiveness of the research, especially considering the unsolved importance of the topic.
We already knew intelligence was good, we are just somewhat more confident in the national results now. Policy implications from my POV are simple: boost your national intelligence. This can be done by 1) reversing dysgenic fertility patterns, 2) high skill immigration or at least avoid low skill immigration, 3) biotech eugenics, i.e., embryo selection and genetic engineering. I think govts not subsidize the latter. I'm not a fan of sterilization programs.
Great work. I am surprised that the socialist/capitalism variable has such a small effect if any. As just eyeballing a IQ vs GDP per capita graph. The two things that stand out are communist and former communist perform below trend and oil producing countries perform above trend.
So is capitalism = "Degree of Capitalism" and Socialist = "Number of Years Open Economy" from Jones, G., & Schneider, W. J. (2006) on "Table 1 Variables from Sala-i-Martin (1997a,b)" or maybe was socialist was "Government consumption share" from Jones, G., & Schneider, W. J. (2006 ) on "Table 5 Growth variables passing the robustness test of SDM (2004), in rank order"?
And would a Dummy variable of whether the country is socialist/communist or has ever been for greater than 3y have a bigger impact on growth in the model?
I am also surprised, but the numbers are what they are. The approach we used with a continuous variable should be stronger than a simple dummy approach. Beware that a dummy approach will essentially function as a proxy for East Europe cultural/genetic factors too. Maybe ask over at George's post.
Hi Tim, a list of all variables and their origin are given in the online appendix. https://osf.io/4x38f/
I was also surprised to see free market indexes perform so poorly. To give free markets the best chance possible, in the extra variable section we replaced the capitalism measures that SDM use with the Fraser Institute's index and it still performs poorly. The Fraser institute measure was used in the BI dataset in all tests.
We discuss reasons for this result in much more depth in the limitations section. But briefly I suggest that free markets underperform because 1) measures of institutions are poor and noisy 2) Selection bias into the sample removes the most socialist countries 3) BMA is an extremely harsh test which only the very very best can pass. IQ is certainly not the whole story!
Is the inverse correlation between 1960's GDP and GDP growth just an artifact because Asia was relatively underdeveloped compared to Europe, at the time, but also capable of faster growth due to higher average IQ? I see you did an analysis by 20 year rolling windows, how does the correlation change between those?
Have you also tried breaking the analysis down by region? Like, are national IQ differences within a region large enough to show a difference in growth rate if the dataset is restricted to only countries in Europe, or only Asia, or only South America?
1. It's a general and extremely robust finding of the growth literature that there is an advantage of backwardness. This has been found in analyses within regions of the world and within country analyses.
2. We tried a test where we forced regional dummies into every model and the results for IQ were robust to this. I wouldn't want to subdivide the dataset into regions because they already have a small sample size. But this test has been done in other studies. Jones and Schneider tried a test varying the inclusion of OECD countries.
Excellent post. One of several solid economic arguments against Third World immigration. Should send this to a certain group of libertarian economists, although I doubt it would change any of their minds.
I suspect mining/primary resource extraction doing well is a function of the 2000s-2010s commodity supercycle thanks to extraordinarily rapid Chinese industrial expansion/economic growth. If true, then primary resource exporters would be expected to perform well below par in the near future, assuming breakneck Chinese growth is basically over.
Intelligence AND an expanding population are the two necessary keys to growth. Japan has stopped growing, and China soon will due to shrinking populations.
Awesome paper! Growth and development research has undoubtedly been biased in focusing on institutions and geography. Culture has been avoided partly for the absence of objective indicators. Hopefully, this would shed some insight into cultural and behavioral patterns that affect growth rates and incomes. Personal discount rates, financial literacy, and work ethic might also be considered for future research.
However, the strong predictive power of IQ along with the Flynn effect intuitively makes reverse causality seem like a more important factor than has been found. Although it is notoriously hard to pin down. Any idea on what specific percentage can be attributed to reverse causation? Over 10%?
There are some studies that looked at Flynn effect and economic growth. As I recall, their methods were subpar, so this question need to be revisited. Christainsen looked for evidence of reverse causality using countries that get rich by luck due to natural resources, and he doesn't find that they have any higher intelligence (e.g. Saudi Arabia, Emirates compared with poor neighbors).
Luck factors for wealth seem like a useful control. Although I'm not sure oil-rich countries have made significant enough public interventions and investments to rule out reverse causality.
Looking at certain East Asian countries (Korea, Japan) that have made these interventions and have experienced growth in both IQ and GDP may shed some light on the issue if certain controls are used (different regions making different amounts of investment and experiencing different rates of growth in both factors).
While this correlation is pretty intuitive I think something like "standard of living" would be more descriptive rather than "GDP growth". GDP isn't the best metric for the economy because it doesn't distinguish between financial assets and real earned wealth.
The US has been "growing" since the 80s if you focus on GDP only. However the real economy sans financial assets has been stagnant implying the economy hasn't really grown. I think high national IQ represents worker capital like you said and therfore potential for a high standard of living.
I agree with GDP criticism, but that's what people use due to data availability going back many decades. I don't have any alternative to use here.
IQ is not intelligence. Defining intelligence by IQ gives circular reasoning. b.-Not all countries have IQ records, especially in Africa, Asia, South America. And they are not massive, on the CIA map they took the IQ of 250 people for a country of 190 million people. It is not significant. And it's not the "national" IQ. That means that study is fundamentally flawed and the data is unreliable.
c- If it were for intelligence, the Mayans, Aztecs and Olmecs had an advanced civilization and the pre-Incas when in Europe were in the stone age. The Nubians who were black had a similar civilization to the Egyptians 5000 years ago. Civilization was born in the tropics, not in the north by brown people. Not white. That contradicts this study, since we all have the same intelligence. If not, why didn't civilization begin in Europe?
https://youtu.be/bEBUr_suntw
Excellent post. I read Garett Jones' book "Hive Mind". It is a good book. There are people like Jeffery Sachs, Francis Fukuyama, and James Robinson (of "why nations fail" fame) that study why some countries do better than other. But of course they shy away from topics like IQ. Instead they use euphemisms like "culture" and education in a lot of their work.
High intelligence culture is surely a good proxy for high intelligence. :o)
I would be interested to read your policy recommendations based on these findings. The reason this type of research is taboo is obviously that people find it offensive and imagine that the policy implications can only be terrible, such as mass forced sterilization programs. The first part may not be changeable, but the second possibly could be. If there were less awful policy implications, people might be more willing to tolerate the offensiveness of the research, especially considering the unsolved importance of the topic.
We already knew intelligence was good, we are just somewhat more confident in the national results now. Policy implications from my POV are simple: boost your national intelligence. This can be done by 1) reversing dysgenic fertility patterns, 2) high skill immigration or at least avoid low skill immigration, 3) biotech eugenics, i.e., embryo selection and genetic engineering. I think govts not subsidize the latter. I'm not a fan of sterilization programs.
The 'conclusion' section of the paper discusses the policy implications and what the ethical and consensual approach could be.
Great work. I am surprised that the socialist/capitalism variable has such a small effect if any. As just eyeballing a IQ vs GDP per capita graph. The two things that stand out are communist and former communist perform below trend and oil producing countries perform above trend.
So is capitalism = "Degree of Capitalism" and Socialist = "Number of Years Open Economy" from Jones, G., & Schneider, W. J. (2006) on "Table 1 Variables from Sala-i-Martin (1997a,b)" or maybe was socialist was "Government consumption share" from Jones, G., & Schneider, W. J. (2006 ) on "Table 5 Growth variables passing the robustness test of SDM (2004), in rank order"?
And would a Dummy variable of whether the country is socialist/communist or has ever been for greater than 3y have a bigger impact on growth in the model?
I am also surprised, but the numbers are what they are. The approach we used with a continuous variable should be stronger than a simple dummy approach. Beware that a dummy approach will essentially function as a proxy for East Europe cultural/genetic factors too. Maybe ask over at George's post.
Hi Tim, a list of all variables and their origin are given in the online appendix. https://osf.io/4x38f/
I was also surprised to see free market indexes perform so poorly. To give free markets the best chance possible, in the extra variable section we replaced the capitalism measures that SDM use with the Fraser Institute's index and it still performs poorly. The Fraser institute measure was used in the BI dataset in all tests.
We discuss reasons for this result in much more depth in the limitations section. But briefly I suggest that free markets underperform because 1) measures of institutions are poor and noisy 2) Selection bias into the sample removes the most socialist countries 3) BMA is an extremely harsh test which only the very very best can pass. IQ is certainly not the whole story!
Is the inverse correlation between 1960's GDP and GDP growth just an artifact because Asia was relatively underdeveloped compared to Europe, at the time, but also capable of faster growth due to higher average IQ? I see you did an analysis by 20 year rolling windows, how does the correlation change between those?
Have you also tried breaking the analysis down by region? Like, are national IQ differences within a region large enough to show a difference in growth rate if the dataset is restricted to only countries in Europe, or only Asia, or only South America?
1. It's a general and extremely robust finding of the growth literature that there is an advantage of backwardness. This has been found in analyses within regions of the world and within country analyses.
2. We tried a test where we forced regional dummies into every model and the results for IQ were robust to this. I wouldn't want to subdivide the dataset into regions because they already have a small sample size. But this test has been done in other studies. Jones and Schneider tried a test varying the inclusion of OECD countries.
Excellent post. One of several solid economic arguments against Third World immigration. Should send this to a certain group of libertarian economists, although I doubt it would change any of their minds.
I suspect mining/primary resource extraction doing well is a function of the 2000s-2010s commodity supercycle thanks to extraordinarily rapid Chinese industrial expansion/economic growth. If true, then primary resource exporters would be expected to perform well below par in the near future, assuming breakneck Chinese growth is basically over.
Intelligence AND an expanding population are the two necessary keys to growth. Japan has stopped growing, and China soon will due to shrinking populations.
Seems intuitive, but fertility failed to predict in the models.
Awesome paper! Growth and development research has undoubtedly been biased in focusing on institutions and geography. Culture has been avoided partly for the absence of objective indicators. Hopefully, this would shed some insight into cultural and behavioral patterns that affect growth rates and incomes. Personal discount rates, financial literacy, and work ethic might also be considered for future research.
However, the strong predictive power of IQ along with the Flynn effect intuitively makes reverse causality seem like a more important factor than has been found. Although it is notoriously hard to pin down. Any idea on what specific percentage can be attributed to reverse causation? Over 10%?
There are some studies that looked at Flynn effect and economic growth. As I recall, their methods were subpar, so this question need to be revisited. Christainsen looked for evidence of reverse causality using countries that get rich by luck due to natural resources, and he doesn't find that they have any higher intelligence (e.g. Saudi Arabia, Emirates compared with poor neighbors).
https://www.sciencedirect.com/science/article/pii/S016028961300113X?casa_token=sLOXE8D3A0MAAAAA:5NJBHO_hSQUkASywVQTTEYscEnYr7EB4uvITKdhZDuf70TPH_5h1uFsrioPdgeTr15UJCVHzSL4
Luck factors for wealth seem like a useful control. Although I'm not sure oil-rich countries have made significant enough public interventions and investments to rule out reverse causality.
Looking at certain East Asian countries (Korea, Japan) that have made these interventions and have experienced growth in both IQ and GDP may shed some light on the issue if certain controls are used (different regions making different amounts of investment and experiencing different rates of growth in both factors).
Luck factors are included as controls, that's what mining GDP and so on are.