«Abstract This chapter uses statistical techniques to assess whether there is evidence that Africa’s slave trades had a detrimental impact on ...»
Chapter 5. Shackled to the Past: The Causes
and Consequences of Africa’s Slave Trades
This chapter uses statistical techniques to assess whether there is evidence that Africa’s
slave trades had a detrimental impact on long-term economic development. This is done
by ﬁrst constructing estimates of the number of slaves taken from each region of Africa
between 1400 and 1900. The estimates are constructed by combining data on the number
of slaves shipped from African ports with data from historical records reporting the ethnic identities of slaves taken from Africa. Using the constructed data, it shown that the parts of the continent from which the largest number of slaves were taken in the past are the parts of the continent that are the poorest today. This relationship is found to be extremely robust. It remains even when other important determinants of economic development are taken into account.
This relationship can be interpreted a number of ways. One interpretation is that it shows that the slave trades had an adverse eﬀect on Africa’s long-term economic develop- ment. An alternative interpretation, however, is that the parts of Africa from which the largest number of slaves were taken in the past were initially the least developed. And because these characteristics persist today, these parts of Africa continue to be the least developed. Therefore, we observe that the parts of Africa that exported many slaves in the past are also poor today, even though the slave trades did not cause these areas to become underdeveloped. This alternative explanation is tested in the data by examining 1 whether it was in fact the initially least developed parts of Africa that exported the great- est number of slave. Consistent with the historical evidence, the data suggest that the parts of Africa that were initially the most developed, not the least developed, supplied the largest number of slaves. This evidence provide strong evidence against the second interpretation, and instead supports the ﬁrst interpretation. It is also shown that addi- tional statistical tests, using instrumental variables, also provides additional support for the slave trades having a causal adverse eﬀect on economic development within Africa.
2 Introduction Africa’s history is intimately connected with slavery. The continent has experience four large slave trades, all of which date back at least to the mid-ﬁfteenth century. The oldest of the slave trades, the trans-Saharan, Red Sea, and Indian Ocean slave trades, all date back to at least 800 ad. During these trades, slaves were taken from land south of the Saharan desert, inland of the Red Sea, and inland of the coast of Eastern Africa, and were shipped to Northern Africa and the Middle East. The largest and most studied of the slave trades is the transAtlantic trade slave trade, where beginning in the ﬁfteenth century slaves were shipped from West Africa, West Central Africa, and Eastern Africa to the European colonies in the New World. Although the Atlantic slave trade was the shortest in duration, it was the largest and most penetrating of the four slave trades. Between the ﬁfteenth and eighteenth centuries, upwards of 12 million slaves were taken from the continent of Africa. The total number of slaves shipped during this same time period in the other three slave trades is somewhere around 6 million. In total, nearly 18 million slaves were shipped in the four slave trade over this four hundred year period.2 Given the sheer magnitude of the slave trades it is natural to ask what eﬀect, if any, the slave trades had on African societies. This is an old and much debated question in the African history literature. A number of authors, dating back to at least the writings of Basil Davidson and Walter Rodney, argue that the slave trades had a signiﬁcant adverse impact on the political, social and economic development of Africa.3 For example, in his book Slavery and African Life, Patrick Manning argues that: “Slavery was corruption: it involved theft, bribery, and exercise of brute force as well as ruses. Slavery thus may be seen as one source of precolonial origins for modern corruption.”4 Along similar lines, Joseph Inikori argues that the long-term consequences of Africa’s slave trades was to “alter the direction of the economic process in Africa away from development and towards underdevelopment and dependence.”5 Recent research has examined the impact of the slave trades on speciﬁc ethnic groups.
trades had on the institutional and social structures of African societies. They show how the external demand for slaves caused political instability, weakened states, promoted political and social fragmentation, and resulted in a deterioration of domestic legal institutions.6 The view of others, such as John Fage and David Northrup, is that the slave trades had little aﬀect on the subsequent socio-economic development of Africa.7 David Northrup, examining the eﬀects of the slave trade in Southeastern Nigeria, concludes that “while it is true that the slave trade was cruel and produced a climate of fear and suspicion, its social and economic eﬀects which can be measured were surprisingly benign.”8 These diﬀerences in opinion are not surprising. Even direct observers of the slave trades had very diﬀerent views of the eﬀects that the slave trades were having on African societies at the time. For example, while the English slave trader, Archibald Dalzel felt that African societies were unaﬀected by the slave trade, the explorer and missionary, David Livingstone argued that the slave trade had a devastating impact on African societies.9 This chapter attempts to shed light on this issue by using statistical analysis to examine the relationship between the severity of the slave trades and subsequent economic performance for diﬀerent parts of Africa. This is done by ﬁrst constructing estimates of the number of slaves taken from each region of Africa between 1400 and 1900. The estimates are constructed by combining data of the number of slaves shipped from each African port or coastal region with data from historical documents reporting the ethnic identities of slaves shipped from Africa. The construction of the slave export estimates builds on the vast empirical literature that has evolved from the research of African historians over the past four decades. Because data on current economic performance, such as per capita Gross Domestic Product (GDP), are only available at the national level, the statistical tests performed in the chapter use current countries as the unit of observation. Because of this, when estimates of the number of slaves taken from diﬀerent parts of Africa are constructed, a ‘part’ is deﬁned as the portion
pletely arbitrary, particularly from a historical perspective, the limited availability of current economic data necessitates the use of the modern nation state as the unit of analysis.
An issue that arises when using modern countries as the unit of analysis is that they are diﬀerent sizes. Therefore, variation in the number of slaves taken from diﬀerent countries will reﬂect, and least to some degree, diﬀerences in country size. Because of this, the constructed country-level slave export measures are ‘normalized’ to take into account diﬀerences in country size.
The logic of the statistical tests is as follows. If the slave trades are partly responsible for Africa’s current underdevelopment, then one should observe that the parts of the continent from which the largest number of slaves were taken in the past are also the parts of the continent that are the poorest today. The tests performed in this chapter examine whether this pattern is observed in the data. The results conﬁrm that the areas from which the most slaves were taken are indeed the parts of Africa that are the poorest today. As will be shown, this relationship is extremely strong, and it remains even when other important determinants of economic development, such as climate, geography, natural resource endowments, and past colonial experience, are taken into account. Although these statistical correlations provide evidence that the slave trades adversely aﬀected Africa’s economic development, this evidence is still not conclusive. The reason is that it may have been the case that the parts of Africa from which the largest number of slaves were taken were initially the most underdeveloped.
Because these characteristics persist today, these parts of Africa continue to be relatively underdeveloped. Therefore, one would observe that the parts of Africa that exported many slaves in the past are also poor today, even if the slave trades did not cause these areas to become underdeveloped. This alternative explanation is tested in the data by examining whether it was in fact the initially least developed parts of Africa that engaged most heavily in the slave trades. Consistent with the historical evidence, the data suggest that the parts
number of slaves.
Although this line of inquiry is diﬀerent from previous historical research that examines the impacts of Africa’s slave trades, the results presented here complement the evidence from these previous studies. For example, the macro-statistical perspective of this study complements more micro-level historical case studies, such as Walter Hawthorne’s analysis of the impact of the slave trade on the Balanta, or Andrew Hubbell’s study of the eﬀects of the slave trade in the region of Souroudougou.10 If the slave trades were detrimental for subsequent social and economic development, then these eﬀects should be observed both at the micro level, when looking at speciﬁc ethnic groups during speciﬁc time periods, and at the macro level, when looking at broad patterns across the whole African continent over a longer time horizon. It may be that the slave trades had very speciﬁc impacts in some places during certain periods of time, but that these are not general eﬀects present across a wide cross-section of societies.
Evidence from the broader, more macro perspective presented here can shed light on how generalizable speciﬁc examples are. The use of macro-statistical analysis also complements broad historical studies that also take a macro perspective and examine the larger impacts of the slave trades within Africa. Examples of studies of this type include Paul Lovejoy’s Transformations in Slavery and Patrick Manning’s Slavery and Occidental Life.11 This study can be seen as an extension of this line of research that simply applies more formal statistical techniques to examine the economic impacts of Africa’s slave trades.
Estimates of the Number of Slaves from African Countries Construction Procedure The analysis of this chapter builds on a long empirical tradition in the African history literature. The seminal work is Philip Curtin’s (1969) The Atlantic Slave Trade: A Census, which used data available at the time to provide a detailed description and comprehensive analysis
Curtin’s publication in 1969, a very impressive amount of additional information has been collected and analyzed by African historians. The most recent and most extensive eﬀorts are the Trans-Atlantic Slave Trade Database, which was developed by David Eltis, Stephen Behrendt, David Richardson, and Herbert Klein (1999), as well as the Louisiana Slave Database and the Louisiana Free Database, constructed by Gwendolyn Midlo Hall (2005).13 Another notable contribution to this literature is Patrick Manning’s of computer models to generate simulations of the estimated demographic impacts of Africa’s slave trades. The results were presented in a series of journal articles and in his book Slavery and African Life, which was published in 1990.14 The present analysis extends this line research by using the wealth of available data to construct estimates of the number of slaves that were taken from the diﬀerent parts of Africa.
Then, the statistical relationship between the number of slaves taken in the past and current economic performance is examined.
The data used to construct the slave export estimates can be grouped into two categories.
The ﬁrst category includes data that report the total number of slaves exported from each port or region in Africa. For the trans-Atlantic slave trade, the data are from the updated version of the Trans-Atlantic Slave Trade Database, which records 34,584 voyages from 1514 to 1866. These data are gathered from documents and records located around the world.
In most European ports, merchants were required to register their ships, declare the volume and value of goods transported, pay duties, and obtain formal permission to leave the port.
Therefore, for each ship and voyage, typically, there exists a number of diﬀerent registers and documents. In the database, 77% of the trans-Atlantic slave voyages after 1700 have shipping information from more than one source. Speciﬁc voyages are documented in as many as sixteen diﬀerent sources. The average number of sources of data for each voyage is six. According to the authors’ estimates, the database contains 82% of all trans-Atlantic
Slave Trade Database is in 1526, decades after the beginning of the trans-Atlantic slave trade.