Correlates of post-colonial development in Africa

Botswana (GDP per capita (PPP): $12,100 ) and Mauritius ($12,400), are doing a lot better than a lot of other sub-Saharan nations. Why? A draft paper “Constitutions, Private Property, and Economic Growth in Africa: Decolonization Processes and Post-colonial Reforms Matter” (pdf) argues that:

…legal history indicates that most African countries repealed British-inspired constitutional safeguards against property expropriation after independence while Botswana and Mauritius kept all constitutional safeguards… This paper provides empirical evidence showing that securing private property was a key component of economic development.

Though, one should also note that Mauritius is 68% Indian – and so hardly qualifies as African. And those countries are just run better in general.


A 2010 paper, “The Causal History of Africa: A
Response to Hopkins” comments on some aspects of colonialism:

Bertocchi and Canova (2002) find that the identity of the colonizer is significant in a cross-country growth regression, while Agbor, Fedderke, and Viege (2009) similarly show that British colonies have grown faster than French ones. Grier (1999) finds that colonies that were held longer are performing better today; Olsson (2009) obtains the same results using democracy as an outcome.

As the author of the first paper points out:

Comparative studies on British and French decolonization in Africa indicates that the British decolonization was done gradually based on
experiences in South Asia whereas the French decolonization was done more in haste after the costly independence wars in Indochina and Algeria, which had different consequences in imposing legal systems

Another paper arguing that British colonies are better off is a 2010 paper “Comparing British and French Colonial Legacies: A Discontinuity Analysis of Cameroon” (pdf). It’s an interesting case because it was originally Germany, but after WWI was split between Britain and France, and then re-united at independence. They conclude:

Taking advantage of the artificial nature of the former colonial boundary, we use it as a discontinuity within a national demographic survey. We show that rural areas on the British side of discontinuity have higher levels of wealth and local public provision of improved water sources. Results for urban areas and centrally-provided public goods show no such effect, suggesting that post-independence policies also play a role in shaping outcomes.

[A] limitation of our results is that we cannot know by what mechanism British colonialism causes superior outcomes.

[But before any British start feeling too satisfied with themselves, the 2010 paper “Direct versus Indirect Colonial Rule in India: Long-Term Consequences” (pdf) found that those regions of India ruled indirectly are doing better now.]


Institutions in African history and development: A review essay” (pdf) 2010:

Nunn (2008) shows that the countries that exported the most slaves are poorest today. Nunn and Wantchekon (2008) fi nd that the slave trade also produced lower levels of trust among the ethnic groups that were most a ffected.

Gennaioli and Rainer (2007) show that pre-colonial state centralization is positively correlated with modern GDP; they posit that rulers of more centralized pre-colonial states were better able to extract public goods from colonial authorities. Bolt and Smits (2010) add local structures to this analysis, and show that countries whose pre-colonial societies had well developed community hierarchies and were outward looking are better governed in the present.

A 2011 paper: “Divide and Rule or the Rule of the Divided? Evidence from Africa.”

Our analysis shows that political complexity before the advent of European colonizers correlates significantly with contemporary development, even when we account for national policies and other country-specific features. This correlation does not necessarily imply a causal relationship because one cannot rule out the possibility that other ethnic characteristics and hard-to account for factors related to land endowments or the ecology drive the association between pre-colonial ethnic institutional traits and development. Yet the positive association between historical institutions and luminosity prevails numerous permutations. First, it is robust to an array of controls related to the disease environment, land endowments, and natural resources among others. Second, regressing luminosity on a variety of alternative pre-colonial ethnicity-specific economic and cultural traits reported by Murdock (1967), we find that political centralization is the strongest correlate of regional economic development. Third, we find that the positive correlation between ethnic historical political complexity and regional development obtains across pairs of adjacent ethnic homelands where groups with different pre-colonial institutions reside. Thus, although we do not have random assignment in ethnic institutions and it is therefore hard to establish causality, the results clearly point out that, unlike national institutions, traits manifested in differences in the pre-colonial institutional legacy of each ethnic group matter for contemporary African development.

They suggest the reason for this is the limited penetration of European influence.


Finally, a paper that argues the geographical pattern of warfare in Africa from 1400 – 1700 correlates to contemporary patterns: “The Legacy of Historical Conflict Evidence from Africa” (2012): “We find robust evidence that patterns of conflict after countries in Africa gained independence are correlated with having had more historical conflicts within their borders. We also find some evidence supporting the view that the mechanism at work may be a diminution intrust, a stronger sense of ethnic identity and a weaker sense of national identity.”


In my earlier post on the origins of inequality, I also discuss the importance of institutions and demographics for current outcomes, but more globally and over a longer time period.

Racial stratification in America’s future

The demographic makeup of the U.S. is changing rapidly, but the human capital of the populations that make it up is not.

A recent report (pdf) by United for a Fair Economy looked at socio-economic racial differences from 1980 until the present, and projected those trends into 2042, when the white population is expected to dip below 50% of the American population.

First, this is how fast the change is happening:

However, though the population share of these populations is booming, their relative wealth is not – meaning that this America of the future, if these trends hold, is going to be even more racially stratified than it is now. There has been remarkably little relative change, or even, in the case of Hispanics, negative change (obviously due to immigration), in some indicators:

The study summarizes what we can expect in 2042 if previous trends hold:

If we continue on the path that we have been on since 1980, in 2042 the median Black family will still earn only about 61 cents for every dollar of  income earned by the median White family, while Latinos will earn just 45 cents for every dollar of White median-family income. Meanwhile, Black  poverty rates will still be close to double that of Whites, and poverty for Latinos will have grown to be more than two-and-a-half times that of Whites. Blacks will face an unemployment rate that is 1.8 times that of Whites, while Latinos will face an unemployment rate that remains roughly 1.5 times that of Whites.

Prison populations will continue to be overwhelmingly Black and Latino. If the trends since 1980 continue, nearly 5 percent of the Black population and close to 2 percent of the Latino population will be in prison in 2042.

Remember, that in 2042, blacks and hispanics are projected to make up about 40% of the population. Consider what that means for the health of American society.

For another perspective on the future of racial stratification, we can look at racial stratification in marriages. The wealth and education of mixed race couples is generally intermediate to that of their respective races. Given that non-white populations are dramatically increasing, this suggests that we may be heading towards less of a black-white dichotomy, and more of a latin america style colour gradient in socio-economic status, with Eurasians at the top, and blacks at the bottom:

I wrote another post, on South Africa, looking at things from the similar perspective of demographic change on economic indicators.

Social Mobility

There’s been talk of a new paper by Gregory Clark measuring social mobility in England over 200 years by using rare surnames (“Are there Ruling Classes? Surnames and Social Mobility in England, 1800-2011“). Here is a graph from the paper:

The abstract:

Using rare surnames we track the socio-economic status of descendants of a sample of English rich and poor in 1800, until 2011. We measure social status through wealth, education, occupation, and age at death. Our method allows unbiased estimates of mobility rates. Paradoxically, we find two things. Mobility rates are lower than conventionally estimated. There is considerable persistence of status, even after 200 years. But there is convergence with each generation. The 1800 underclass has already attained mediocrity. And the 1800 upper class will eventually dissolve into the mass of society, though perhaps not for another 300 years, or longer.

Burakumin and Koreans in Japan

Often cited in the IQ debates are the Burakumin (and sometimes Koreans as well) of Japan. For example, in Race and intelligence: separating science from myth (2001), Ogbu says

Time magazine had the following to say about the Buraku-Ippan gap in IQ test scores in 1973:

“…their children test 16 IQ points lower than other Japanese. (Remarkably similar to the average 15-point difference between U.S. Blacks and Whites…”

As voluntary minorities, the Burakumin in the United States perform well. Or, to put it differently, there is no evidence that they do less well on IQ tests and in school than other Japanese immigrants. In fact, the only study of Japanese immigrant performance in the United States that identified them indicates that they do slightly better in school than other Japanese immigrants (Ito, 1967).

One issue we should dispose of immediately is that there is not actually any evidence Burakumin do any better in the U.S. As Jason Malloy points out:

There is no data for Burakumin in the US. False claims about US IQ data have mutated second-hand from John Ogbu who claimed a study showed that the Baraku immigrants here “do slightly better in school than the other Japanese immigrants”. The book chapter Ogbu references for this claim (Ito 1966) however, is by a pseudonymous author who relied strictly on gossip from non-outcast Japanese communities in California to surmise how the outcasts here might be performing. The author’s informants believed the US outcasts were more attractive, more fair-skinned, and made more money. Though– as a testament to Ogbu’s immaculate scholarship– the author reported no gossip about how these Burakumin performed in school.
* Ito, H. (1966) Japan’s outcastes in the United States. In G.A. deVos and H. Wagatsuma (eds.), Japan’s Invisible Race. Berkeley: University of California Press.”

The only IQ tests done on Burakumin were decades ago, but they do show, as noted above, a gap similar to that between blacks and whites in the U.S (DeVos 1973; Ddevos & Wagatsuma, 1967). The academic achievement gap has persisted. Burakumin do worse academically, and enter college and high school at lower rates:

Burakumin IQ scores. From Ogbu (2001), in "Race and intelligence: separating science from myth"

Ikeda (2005) "Ethnicity, race, and nationality in education: a global pespective"

High School entrance rates - Burakumin: blue diamonds, Japanese majority: pink squares. (source:

College entrance rates - Burakumin: blue diamonds, Japanese majority: pink squares (source:

There have not been any IQ tests on Koreans in Japan, but they show a similarly low achievement academically, despite Koreans in Korea and the Western world doing very well. Koreans living in Japan originates with Japan’s colonization of Korea and the sometimes forced, sometimes voluntary movement of people to fill labour shortages. Like Burakumin, Koreans have faced general discrimination. As we can see below, Koreans do even worse than Burakumin in qualifying for higher education:

Percent of each group that are qualified by grades for higher institutions. From Ogbu (2001), in "Race and intelligence: separating science from myth"

Koreans and Burakumin also show other persistent social pathologies that make them reminiscent of blacks, or other unsuccessful minorities. As of 1995, Burakumin earn 60% of the national average, single parents are twice as common, welfare dependency is 7x the national average, one study (in the 60s – the issue is very sensitive) found Burakumin youth 3x more likely to be arrested. Burakumin and Koreans are known to dominate the organized crime business as well – the Yamaguchi Gumi, the biggest gang (at least in 1995), was 75% Burakumin and Korean.

These problems persist despite the govenment spending substantial amounts of money trying to improve Burakumin social and educational conditions (Shimahara, 1991).

Given that Burakumin are ethnically Japanese and suffered discrimination on the order of centuries, and Koreans only arrived in Japan in the last century and have been successful in all other parts of the world, these are unlikely to be genetically caused differences, and yet they show very similar features to the putative genetic differences between blacks and whites.

The Origins of Inequality

The unequal global distribution of technology and wealth between the ‘North’ and the ‘South’ began a long time ago.

We all know that the early development in agriculture and civilization began in Eurasia – but it goes back much earlier than that.

At a somewhat atemporal level, toolkits of hunter-gatherer have been shown to increase in size and complexity with latitude (Oswalt, 1976) – the driving cause seeming to be the risk of resource failure (Collard et al., 2005).

Developments necessary for cold, risk of resource failure… whatever the cause, Foley (1987) writes: “although there is a general and global technological development during the Pleistocene, it is in high latitudes that it is most marked; in parts of the tropics the artefacts remained simple.”

Hoffecker (2005) notes that evidence of “technological complexity and creativity in the archeological record is broadly coincident with and presumably related to traces of creativity in art, music, ritual, and other forms of symbolism. The pattern of modern human technology is part of a larger package of behavior (sometimes referred to as “behavioral modernity”) that emerges with the appearance of industries in Eurasia classified as Upper Paleolithic, but has deeper roots in the African Middle Stone Age.” He adds that, while traditionally the early technological innovations of Eurasians have been attributed to a need to adapt to colder climates, some of them, such as notation and musical instruments, cannot be explained in that way.

Technological adoption in 1000 BC vs modern per capita income

Comin et al. (2008) asked “Was the Wealth of Nations Determined in 1000 BC?” They collected data on the technology adoption of countries in 1000 BC, 0 AD, 1000 AD and 1500 AD, and compared to contemporary per capita income. While of course the strongest correlation was 1500 AD, there was nevertheless a definite correlation going back as far as 1000 BC. As mentioned above, this technological precociousness is in fact much older than that, preceding even agriculture.

The conditions which allowed agriculture to flourish and spread are well known from Jared Diamond’s Guns, Germs, and Steel. Following Diamond, Hibbs & Olsson (2004) quantify those types of geographical and biogeographical traits ( which, even if not as complete an explanation as Diamond would like them to be, they are certainly important) and compared them to time of transition to agriculture. Of course, these were factors that were especially beneficial for Eurasian populations. The authors also compared the scores on those initial environmental traits to modern GDP, results shown to the right.

So the inequality of regions goes back a long way. But while regional geographical/biogeographical characterisitcs are clearly of great importance in determining the prospects of the people who live there, we can’t forget the importance of the populations themselves. In other words: sometimes it is the population that matters, not the region.

This is best illustrated in events over the last 500 years. By 1500 AD, there was a quite a bit of disparity in terms of wealth, technology, and social organization between world populations.

[In fact, by that point people had even become, to some degree, more biologically adapted to their own environment – ie. Eurasians had immunities to the diseases that plagued their cities, Andean and Tibetan peoples had adapted to their high altitudes, Inuit to the cold Arctic environment, Africans to the tropical disease burden, etc. etc. These factors themselves played an important role in the ability of modern groups to succeed in different environments. Europeans had much difficulty in governing, let alone settling, tropical nations due to disease, whereas they flourished in more temperate colonies – especially those, like in the Americas, where the natives helpfully died off in the millions due to lack of immunity to Eurasian diseases, and Europeans could replace them fairly quickly, before they had time to rebound. The point of this aside is that, while over a large time scale environment may shape human societies, over a shorter time scale, and with repercussions that can have great long term consequences, the specific populations involved are of great importance – Europeans came to replace the populations in some parts of the world, and not in other parts, often because they were suited towards some environments, and not others, biologically.]

In the same way, the long history of cultural and technological developments of some populations, and not others, needs to be understood in order to explain why two populations, though now living in the same country, or even same neighbourhood, can have such an unequal distribution of wealth, education, social pathologies, etc. It is because populations carry something with them, something that has shaped them into a certain mold by a very long history – and that other populations do not share, and can not come to share easily. In my post on the African marriage system and the high rate of STDs among blacks, I discuss another important instance of a population carrying with it behaviours that cause to have a certain pattern of STD infection, even when in a new country, next to populations that show very different patterns.

Since 1500 AD, there has been a dramatic moving of populations, illustrated by this graphic from Putterman and Weil (2010):

Putterman and Weil calculated, for each population in 1500 AD, its early developmental history in terms of how long ago it had adopted agriculture and state level organization, and then they traced the movement of these populations around the world and looked at their current level of success. They find that it is the “history of a population’s ancestors rather than the history of the place they live today [that] greatly improves the ability of those indicators to predict current GDP.” And it is correlated “both at the level of country averages and in terms of an individual’s position within a country’s income distribution.” The predicted ordering of inequality based on development history is consistently found across and within countries. The author comments that in explaining variance in today’s country incomes, the state history and agricultural transition variables can each explain about a quarter of the variance. Looking at a sample of 11 countries now containing different populations with markedly different developmental histories, the authors found that in 9 of them the population hierarchies perfectly matched with expectations, and in the other 2, they were broadly consistent with only minor differences.

Technological adoptions in 1500 AD vs modern per capita income

Like for 1000 BC, technology adoption in 1500 AD, was unsurprisingly, still correlated (but more so) to current population success (Comin et al. 2008).

As one of the authors notes in an online article, “78 percent of the difference in income today between sub-Saharan Africa and Western Europe is explained by technology differences that already existed in 1500 A.D. – even BEFORE the slave trade and colonialism.” (I might note here that, by the late middle ages, the estimated per capita income was $1000 in England, which is higher than a number of African countries are now (British Economic Growth 1270-1870. 2010))

Another paper, Foa (2011), also focusses on migration since 1500 AD, but is interested in the fates of the settler colonies of non-Europeans. That is, countries that have, since 1500, become dominated demographicaly by a non-European peoples, and voluntarily so (ie. excluding forced movement via the slave trade). These settler countries include people of Chinese, Indian, and African origins. Chinese settler colonies are: Hong Kong, Singapore, and Taiwan. Indian: Fiji, Trinidad & Tobago, Guyana, and Mauritius. African: Botswana, Swaziland, Namibia, Cape Verde, and Equatorial Guinea.

What is important to remember about all these settler countries is that they all experienced different colonial patterns, being under British, Dutch, Chinese, French, or Portuguese control, and yet it seems to be that it “is not whether a colony was governed by a British or a Portuguese official that determines their extent of state capacity today, but rather whether such colonies were peopled by groups historically accustomed to norms of centralized political organization, capable of maintaining and reconstructing such institutions in the post-colonial era.” As such, it is the Chinese colonies that have done the best, followed by Indian, and Africans doing the least well. I will allow another word on the importance of the population to the author: “when it is argued that the postwar developmental states of Korea or Taiwan trace their origins to Japanese colonial rule, we would be well advised to keep in mind the counterfactual hypothesis, namely as to whether comparable outcomes would have attained had the Japanese not conquered neighbouring lands with similar traditions of bureaucratic development, but instead conquered lands along the Congo or Gambia rivers.”

The following table shows how countries that are now dominated by some particular ethnic settler group does on various measurements: corruption, governmental effectiveness, ‘voice and accountability’ (adherence to democratic norms), and GDP per capita.