On kin groups and employment in Africa

A common feature of African societies is that individuals belong to kin groups which impose reciprocal obligations upon their members. While kin groups are found in other societies, African kin groups are distinctive both by their ubiquity and by the strength of their claims upon members. Section 2 summarises current economic interpretations of kin groups as insurance and transfer institutions which succeed in lowering transactions costs in the traditional, small-scale economy. However, in the modern economy, where large scale production is required, firms must employ multiple kin groups. The resulting employment relations between managers and workers differ from those in which the basic unit is the individual employee. Kin groups will attempt to favour their own members in the assignment of good jobs. A kin group may be restrained in its favouritism either by competition from rival kin groups or by promotion practices imposed by top management. Even though kin group favouritism may be rife, it is not readily observable on conventional labour market data sets. Although we utilise an unusually rich Ghanaian data set, it still does not provide direct evidence enabling us to test whether managers favour workers from their own kin group. However, in Section 3 we set out a model in which kin group favouritism is shown to give rise to a wage premium for the largest kin group. Since our data includes the size of kin groups, and since these vary across localities, we are able to test for kin group favouritism indirectly by investigating whether there is a wage premium purely related to membership of the locally largest kin group. In Section 4 we test for kin group favouritism in Ghana, treating separately private and public firms. We find that in the private sector there is no evidence for kin group favouritism: workers are paid their marginal product. By contrast, in the public sector workers are rewarded for membership of the right kin group, but not for productive characteristics.