7. Exploring the Mechanism of Colonial Rule

Note that this colony was largely uninhabited at the time of colonization.

Richens (2009) finds evidence that indirect rule had worse consequences in the context of fragmented precolonial societies than in politically organized ones. Botswana is a case in point. Precolonial sociopolitical structures and systems of representation of the Tswana chiefdoms could be maintained to a greater extent than in other indirectly ruled colonies because Tswana chiefs, despite collaborating with the British over the whole colonial period, firmly opposed increased colonial control of domestic matters. As a result, institutional constraints to legitimate chiefs were not dismantled, and the tribal assembly could continue to exercise the legislative, executive and judicial powers of the tribe (Lange 2009). This institutional inheritance in the postcolonial period was central for developing checks and balances on political leaders that continued to be accountable to their people, as in precolonial times (Acemoglu and Robinson 2010). Somehow related, Acemoglu et al. (2014a) and Heldring and Robinson (2014) hypothesize that the effect of indirect rule is less pervasive in politically centralized societies. Accordingly, the greater power of traditional elites in precolonial politically centralized states, such as the Asante in Ghana and the Buganda in Uganda, led post-independence leaders to fully dismantle the organizational structures created by indirect rule and marginalize traditional elites—because of the expected difficulty for the former to control such powerful traditional elites. In contrast, in countries lacking centralized polities prior to colonization, whereby chiefs were usually created by the British, chiefs were not abolished and their powers were further institutionalized during the post-independence period. This generated a more intense form of indirect rule, a system of local despotism which effectively impeded state governance and broad-based development (Mamdani 1996). A clear example of the latter is Sierra Leone in which the British, upon their arrival, dismantled the hierarchical structure of the existing indigenous federated states. They were replaced by newly created chiefdoms as a way to weaken any opposition and local resistance via the application of a divide-and-rule policy (Lange 2009).

Acemoglu et al. (2014a) argue that indirect rule strengthened local elites, who were largely unaccountable to their people, and undermined the colonial and postcolonial central state that was non-bureaucratized, lacked a monopoly of violence and a well-functioning fiscal system, thus failing to provide even the most basic public goods. Acemoglu et al. (2014b) provide evidence consistent with these claims for Sierra Leone.

As Lange (2009) points out, an important differentiation within the group of British settler colonies is also in order. Unlike the full-blown settler colonies of North America and Australasia, the settler colonies in sub-Saharan Africa (i.e., Zimbabwe, Kenya and South Africa, and to a lesser extent Swaziland, Botswana and Zambia) featured a much smaller number of European settlers and a much larger number of indigenous people. In these settler colonies that lacked large populations of white settlers, highly extractive labor and land market institutions were established.

Comparing French West Africa with British Africa, Whittlesey (1936/1937, p. 367) notes that “[t]he total white population of the British colonies is about 11,000, that of the French about 31,000. The proportion of traders is notably higher in the British territory because the British employ approximately twice as large a personnel to manage a store as do the French. The ratio of missionaries is probably about the same. The proportion of functionaries is therefore much higher in the French possessions. Obviously more political officers are needed for direct than for indirect government.”

According to Miles (1987), when the French selected chiefs, the ability to speak French or prior service to the colonial administration prevailed over traditional forms of chiefs’ legitimacy to their people. Indeed, progressive erosion of chief’s traditional authority, combined with his unpopular tasks of taxation and recruitment of forced labor and troops, converted him “into an agent of the most hated aspects of French colonial rule” (Crowder 1968, p. 193).

As described by Cogneau and Moradi (2014), while the British entrusted religious missions with education, the French set a network of publicly provided non-religious schools as key to converting Africans into Frenchmen. In French colonies, government-certified teachers were employed, and instruction of government curriculum was conducted only in French. In contrast, missions in British Africa had freedom to run their schools, recruit teachers who could teach in local vernacular languages and adjust the content of the curriculum to local circumstances. In other words, this education system can be thought of as decentralized, whereas the one in French Africa was centralized.

Gallego and Woodberry (2010) assign a key role to competition among missions in accounting for the superior educational outcomes in British Africa.

For instance, if we compare Botswana with Nigeria we can see clearly how endowments determined the extent of indirect rule applied to them. Botswana was relatively sparsely populated compared to Nigeria, with a population density before colonization of 0.28 vs. 23.38. Not surprisingly, the extent of indirect rule was much higher in Nigeria. The number of Africans per European administrator was 56,428 in Nigeria, while 13,193 in Botswana (Richens 2009). If the extent of indirect rule is measured as the ratio of colonially recognized customary court cases over the total number of court cases in 1955, with the latter comprising both customary court cases heard by native chiefs and magistrate court cases handled by British officials (Lange 2004), the extent of indirect rule is 93.4 in Nigeria vs. 42.5 in Botswana.

The small number of observations prevented us from including more variables in the exogenous control set.

These are calculated on the basis of the Anderson and Rubin—AR—(1949) test, the conditional likelihood ratio (CLR) of Moreira (2003) and the Lagrange Multiplier K test of Kleibergen (2005).

Nevertheless, we would like to emphasize that it is impossible to rule out the possibility that the exclusion restriction is violated, particularly in a cross-country regression analysis like the one conducted here. Therefore, the evidence of this 2SLS exercise should be taken as suggestive. In addition, a word of caution about the external validity of this analysis focusing on African colonies is in place. Even though we consider that the evidence observed in Africa reflects a general pattern, more empirical work is needed to validate this hypothesis.