A few months ago I was in Ghana and visited St George’s castle in Elmina, the infamous fort used by the Dutch to hold large numbers of African slaves prior to their ‘export’ to the New World. The Dutch captured the fort from the Portuguese in 1637, and when the slave-trading days were over they handed it over to the British in 1871. St George’s castle is imposing and grim. Its current status as a UNESCO world heritage site is well deserved.
I embarked on a tour with a group of ‘roots-seeking’ Afro-American tourists. We were shown around by an angry young man who clearly disliked European visitors, and laced his stories with gruesome details about how the ‘whites’ had mistreated the ‘blacks’. As the only white person in the group, I felt increasingly uneasy about the relentless flow of stories of rape and murder. The situation deteriorated even further when I tried to correct a small error in the guide’s calculus. Rather than the alleged 10 million Africans exported out of Elmina by the Dutch, the total number was closer to 1 million. ‘… Which is of course still an incredibly big number …’, I added, in an attempt to break the awkward silence that followed. If looks could kill, you would not now be reading these words.
Later that night I mulled over the tour guide’s attitude. Had he discovered that adding a semi-racist edge to his expose was good for business – generating more tips from disgruntled blacks or from guilty whites? Or could it be that he was genuinely upset about these historical events, and that his country was still carrying the burden of slavery so many years after its abolition? Could it be that the shipping off of millions of men and women had left a permanent mark on Africa?
In a recent study, Harvard economist Nathan Nunn tries to explore this issue.1 His first challenge was to pinpoint the areas of origin of exported Africans. Like an archaeologist, Nunn delved deep into ancient shipping records and ethnicity documents. By tallying the results for four distinct slave trade routes between 1400 and 1900, Nunn arrived at estimates of the number of slaves shipped out of each modern African nation state. These numbers varied from over 3.5 million from Nigeria, to none from Botswana and some other countries. Surprisingly, many slaves were caught in landlocked countries. For example, almost a million people were captured in the areas now known as Mali and the Democratic Republic of Congo.
Nunn then proceeds to relate slave exports to modern economic performance, and finds a robust negative correlation. Countries that lost more slaves back then are the ones that have lower incomes today. However, this does not necessarily indicate a causal relation between slave exports and economic performance. It is possible, for example, that regions with weak local institutions between 1400 and 1900 were simultaneously vulnerable to slave raids and also experienced disappointing economic growth. The relation between slave exports and economic performance would then be spurious.
Addressing this concern about an ‘omitted variable bias’, Nunn finds that the negative association between slave exports and income persists. In a tentative and preliminary analysis he identifies two potential channels via which the slave trade may have had a lasting adverse impact. Slave exports must have severely disrupted local societies, and are associated with greater current ethnic fractionalization and reduced state development. In turn, these factors could have impeded economic growth, although their importance vis-à-vis other root causes of underdevelopment is now being debated.2
The Dutch traded slaves until 1814. Many people will probably be as surprised as I was to learn that this historical episode has had a lasting impact on African development. I am unsure whether this has repercussions for the current debate about the moral dimension of development aid. But it will certainly influence my tipping behaviour next time I visit a West African slave fort.