I recently completed my dissertation and thought to share the highlights of it. If anyone is interested in reading the full dissertation please send me an email at nobikil1 [at] Binghamton.edu. A very brief summary is posted below.
In economic development, Africa has lagged behind the rest of the world for many decades, certainly since the early 1960s, when most of today’s African states became independent of European colonial powers. Recent research suggests one reason for Africa’s slow development is its unique experience of the intercontinental slave trade from the fifteenth century through the 1800s, prior to the era of European colonization.
Nunn (2008) examines the relation between GDP per capita of modern African states and the slave trade experience of the ethnic groups now living in the state. He finds that an African country has lower GDP per capita today if the ethnic groups within the country suffered higher slave export intensity – that is, number of slaves exported relative to population – in the precolonial era. He speculates that this is because the production of slaves, which occurred through domestic warfare, raiding and kidnapping, impeded the formation of broader ethnic groups. This may have led to ethnic fractionalization and the weakening and underdevelopment of political structures which hindered the development of postcolonial states. Whatley and Gillezeau (2011) show that, within Africa, areas where demand for slaves was highest are indeed characterized by higher ethnic fractionalization today. Easterly and Levine (1997) show that across countries outside as well as inside Africa, higher ethnic fractionalization is associated with “low schooling, underdeveloped financial systems, distorted foreign exchange markets, and insufficient infrastructure,”(1997, p. 1241) which are in turn associated with low output per worker, at the national level. Easterly and Levine argue this supports “theories that interest group polarization leads to rent-seeking behavior and reduces the consensus for public goods, creating long-run growth tragedies”(1997, p. 1241), and can account for much of Africa’s uniquely poor development. Nunn and Wantchekon (2011) show that Africans belonging to ethnic groups that suffered higher slave export intensity express less trust in relatives, neighbors and local governments today, and suggest this mistrust has damaged modern Africa’s political institutions. All of these studies rely on data from the current era of independent African states. All appear to suggest a channel from pre-colonial slave export intensity to modern-day poverty running mainly through the effect of inter-ethnic group conflict on the quality of post-independence states.
In this dissertation I explore the possibility that the historical experience of the slave trade has reduced per capita GDP in Africa today by retarding the development of human capital through education, specifically the development of literacy. Many studies show that, within Africa and outside Africa, literacy is positively associated with real GDP and other indicators of economic development. Economic theory as well as empirical evidence suggest the relation is at least partly causal, as literacy is an element of human capital. Across countries, differences in average levels of education account for a substantial portion of variations in GDP per worker (Lucas, 1988; Mankiw, Romer and Weil, 1992; Hall and Jones, 1999). Across African countries today, literacy rates are negatively correlated with exposure to the slave trade. Indeed including literacy rates as an explanatory variable in Nunn’s baseline regressions reduces the magnitude and significance of the effect of the slave trade.
Of course, the correlations in the cross country data do not show the direction of causality. If the slave trade damaged Africa’s economic development through any channel, one would expect to see the negative correlation; poor countries cannot provide good schooling. Even if the relationship is causal, it would not be surprising given the common argument that the slave trade damaged African development through the quality of post-independent states. Dysfunctional states also cannot provide good schooling.
My evidence, however, is from the era of European colonization of Africa, after the end of the slave trades but prior to the independence of modern African states. I rely on data from censuses carried out by the British government from the 1920s through the 1950s in the colonies that have become the modern states of Nigeria and Ghana. These colonies contained many different ethnic groups, which had widely differing degrees of exposure to the pre-colonial slave trade. I find that ethnic groups with higher slave trade intensity had lower literacy rates during the colonial era. Using data from recent years, I find that the relation between slave trade intensity and ethnic groups’ literacy persists in contemporary Nigeria. The relationship holds controlling for many confounding factors such as the presence of missionaries, religion, disease environment, the nature of the ethnic groups’ socio political institutions in the precolonial era, and indicators of groups’ economic development in the colonial era. The relation is not related to inter-group conflict in any obvious way, as it holds across small geographic areas that were each dominated by a single ethnic group.
My evidence suggests that the relation between slave-trade history and literacy in Africa today is indeed at least partly causal. Perhaps more importantly, it shows that the slave trades affected current economic development through channels distinct from inter-ethnic group conflict or the quality of formal state institutions. Within British colonial Nigeria and Ghana, the quality of formal state-level institutions was the same for all ethnic groups. The relation between precolonial slave intensity and colonial literacy must reflect something more basic, operating at the level of families or the informal institutions of villages and towns.
I speculate that the relation can be explained by a hypothesis that experience of the slave trade undermined the development of social capital within ethnic groups. Social capital is defined by Putman (1993, pg. 167) as “the features of social organization, such as trust, norms and networks, that can improve the efficiency of society.” Coleman (1988), Knack and Keefer (1997), and Nahapiet and Ghoshal (1998) show that higher levels of social capital are associated with higher levels of human capital. I examine this hypothesis in chapter 2. Using data on expenditure of cocoa farmers in Western Nigeria from 1952, I show that farmers in townships with higher social spending individually spend more on education. The relationship holds after controlling for various characteristics of the farmers and the townships. This suggests that it is possible that the slave trade affects human capital through its destructive effect on social capital.
Finally I examine the effect of the slave trade on political fragmentation within ethnic groups. Using descriptions of local politics of ethnic groups from Murdock’s 1959 ethnographic atlas, I find that ethnic groups with a higher slave export intensity are more politically fragmented.