The worldwide distribution of intelligence

Lesen Sie die deutsche Version hier.

The study of intelligence is fraught with controversy. This is all the more true for research involving group differences in intelligence, whether these concern classes, nations, sexes, and especially races or ethnicities. The English explorer and anthropologist Francis Galton, known for his pioneering studies on human intelligence, attempted to estimate the average intelligence levels of human populations as early as 1869. However, modern, practically useful intelligence tests were not invented until 1905 in France. The first compilation of actual test data on the worldwide variation in intelligence was made in 1978 by British researcher Richard Lynn, who estimated scores for 23 countries. Lynn published an updated compilation in 1991, and in 2002, he published a book with the first truly comprehensive compilation covering 81 countries, including estimates for the remaining 104 countries based on their neighbors. This book caused a massive stir and was much criticized. Lynn continued to update his compilation every few years until his death in 2023.

His last compilation was published in 2019, but was mainly the work of German researcher David Becker. By this point, a number of competing compilations had been made by different research teams. However, as the German researcher Heiner Rindermann showed in 2007, these different compilations were largely in agreement concerning which countries were smart and which less so, and thus could be aggregated into a single set of best estimates.

The beginning of OECD’s PISA program in 2000 marked a turning point. While many people do not think of PISA tests as intelligence tests, but rather as learning or even school quality tests, research shows that they function well as intelligence tests. This is because essentially any test that requires thinking is actually measuring intelligence to some extent. The process of estimating the intelligence of nations has continued to this day, and there are at least six worldwide compilations of national IQs, that is, estimates of the average intelligence of countries. By integrating them into a single set of best estimates is how the world map in figure 1 was produced.

Censored science

National intelligence research is unpopular among left-wing academics, because most academics are politically left-wing. Hundreds of articles in newspapers, magazines, and blogs have claimed that the research is invalid and should be retracted. This is because the average intelligence scores for most groups are below those of the Northwest-central Europeans. Some academics have spent years trying to get all such work censored. If it is already published, they try to get it retracted, and if not, they block its publication.

Typical criticism concerns the low data quality in poor countries (mainly those in Africa), potential test biases (the tests were usually made by Europeans), and some mistakes that were made in earlier compilations.

The low data quality issue is true, but not specific to intelligence research. Crime rates, GDP data etc. from these countries are also of poor quality.¹ Test bias has been extensively tested in Western countries, and there’s little bias for native speakers who grew up there. It’s less clear for recent immigrants, even using non-verbal tests. There’s not much detailed research on test bias between countries because the underlying data are not available for study. In general, however, since all the tests show about the same patterns despite being different in style and origin, makes it unlikely that this is a big issue.

The anti-scientific approach of some of the critics has not prevented the work from being carried out, but it has prevented the integration of research across disciplines. The point is that the correct approach should not be to retract flawed work, but rather to gather new data, and employ new methods. But critics never do that, their goal is not better estimates, but less discussion of the topic.

Some economists, including some at the World Bank, have made their own compilations of the intelligence of nations. However, they never use that word; instead, they refer to the scores by various more vague terms like “test scores”, “human capital”, or “basic skills”. Likewise, they do not cite Richard Lynn’s pioneering work.

Intelligence is the best predictor for growth

Importantly, however, a few economists have documented the importance of intelligence in predicting the economic growth of countries, and have furthermore shown that increases in years of education do not lead to increases in test scores. In fact, at least two studies have tested intelligence against many other variables that are thought to explain differences in wealth between countries. Both analyses found that intelligence was the best predictor along with being relatively poor to begin with. It is easier to grow the economy when one can easily learn from other countries by copying their technology and industrial processes, but much more difficult when new economic growth depends on innovation. Countries thus tend to catch up to their long-term trendline after suffering a decline due to war or communism (think of the rise of Japan, the Asian Tigers, and China, as well as Germany itself after World War II).

«It is easier to grow the economy when one can easily learn from other countries by copying their technology and industrial processes, but much more difficult when new economic growth depends on innovation.»

Identifying the size of the gaps is one thing; another is investigating their causes. In 1969, American researcher Arthur Jensen published an important study examining the causes of the Black-White intelligence gap, that is, the difference between African Americans and European Americans. This gap has been roughly constant at 15 IQ since it was first measured in World War I. After reviewing essentially the entire literature, he concluded that it was “not an unreasonable hypothesis” that genetic factors had something to do with the gap. They could not be adequately explained by test bias, parental factors, school quality and so on. Furthermore, large-scale government programs in remedial or compensatory education had failed to reduce the gap, or even cause permanent increases in intelligence. The high heritability of intelligence within groups (about 80 percent in adulthood, see the article of Robert Plomin) made it difficult to explain the gaps mathematically in terms of known environmental causes.

Jensen’s study drew immediate and intense hostility from left-wing academics. Jensen continued to research and published on these issues the next four decades, and his 2005 review article was particularly important. It concluded that with the data available as it was at the time, it was likely that genetics explained about 80 percent of the variation between individuals and 50 to 80 percent of the Black-White gap.

Evidence from genetics

Jensen died in 2012, but since then, research on the causes of the intelligence gaps has accelerated due to the emergence of studies finding specific genetic variations that explain intelligence (or education, which is very strongly related to intelligence genetically), called genome-wide association study (GWAS). The Italian researcher Davide Piffer has pioneered the use of genetic predictions made from the GWAS models to examine the worldwide distribution of intelligence. In his research from 2013 to the present day, he has found that the genetic predictions line up fairly well with the observed differences between populations. These genetic predictions are not perfect and do not work as well for non-Europeans as they do for Europeans. However, it is difficult to explain why they should match the global distribution of intelligence so well if genetics do not play a role in explaining these. Furthermore, the same methods are able to explain worldwide variation in height, and can even inform us about the recent evolution of intelligence and other traits based on ancient genomes.

Additionally, in 2019, a team headed by myself published a study that examined whether individual variation in ancestry could explain differences in intelligence scores among American children. While a majority of Americans have almost only European ancestry, some have some mixture of African, European, and Native American ancestry. By comparing the ancestry proportions between the children, we found that these could largely explain the differences. This study has since been replicated in two other datasets with similar results. Furthermore, a study from 2016 used the same approach for nations and subnational units in the Americas and also found that the political units with more European ancestry had higher intelligence scores and better socioeconomic outcomes (wealthier, higher life expectancy etc.).

Of particular interest to many has been the extreme success of the Jews seen since the 1800s. For most of their history, Jews kept to themselves and did not contribute to science or philosophy. However, following the enlightenment, more secular Jewish scholars began to make an impact. By the early 1900s, their impact was extreme. Everybody is familiar with luminaries such as Einstein and Oppenheimer. It is probably safe to say that without Jewish physicists, the US would not have been able to make a nuclear bomb in World War II.

«It is probably safe to say that without Jewish physicists, the US would not have been able to make a nuclear bomb in World War II.»

But what about test scores? Results from mainly American studies have shown high averages for Jews, centered around 110 compared to the British or Northwestern European mean of 100. An initial small study from 2019 using genetically predicted scores found that the Ashkenazi (German Jews) Jews had the highest known average score. This result was later replicated in a large sample. Given the evidence then, it is clear that Ashkenazi jews achieve such high status in science, cultural institutions, and society at large mainly because they have high intelligence. Existing genetic studies, while not conclusive, align with this advantage having genetic causes. Similarly, the economic rise of East Asia (countries inhabited by Chinese, Koreans, and Japanese) aligns with IQ studies of these populations, both in their home countries, and as immigrants abroad. Typically, an average IQ around 105 is found, but it depends on the test used (higher on non-verbal) and local immigration selection patterns.

Moving on from the nations and races, readers might also ask: what about men and women? Researchers’ opinions on this topic range from a small advantage for men to about no difference. It is probably safe to say that the latter is the majority opinion, but among those who have studied the topic themselves, I would guess that the majority opinion is that men are probably slightly smarter, perhaps 2 to 4 IQ. This difference is quite small compared to the large differences between countries, and is even harder to study. There are several reasons for that: First, males grow for a longer time during development, and thus one needs data from representative samples of adults to study it well. However, such studies are much rarer than data from children in school. It is hard to get a truly representative sample of adults to participate in a study that involves taking hour-long tests. The second issue is that men and women differ in many cognitive abilities that aren’t just general intelligence. For instance, men clearly have better spatial rotation skills, while women have slightly better episodic memory. Even the words included in vocabulary tests, and the information asked for in knowledge tests can have large impacts on the results (men know more about history, women know more about health).

The large differences on different tests in different directions make it difficult to statistically estimate the difference in general intelligence alone. According to Richard Lynn’s developmental model, girls have a small advantage in intelligence during childhood, especially as they reach the growth spurt of puberty before boys. By age 15, however, boys have caught up, and by adulthood the gap reaches 2 to 4 IQ.