The substitutability of the economic institution of slave labour has often been assumed as a given. Apart from some capital investment to retrain slaves for a different task, essentially their labour could be substituted for any other form of labour.

This paper questions that assumption by using a longitudinal study of the Graaff-Reinet district on the eastern frontier of South Africa’s Cape Colony. We calculate the Hicksian elasticity of complementarity coefficients for each year of a 22-year combination of cross-sectional tax datasets (1805–1828) to test whether slave labour was substitutable for other forms of labour. We find that slave labour, indigenous labour and settler family labour were not substitutable over the period of the study. This lends credence to the finding that slave and family labour were two different inputs in agricultural production. Indigenous khoe labour and slave labour remain complements throughout the period of the study even when khoe labour becomes scarce after the frontier conflicts. We argue that the non-substitutability of slave labour was due to the settlers’ need to acquire labourers with location-specific skills such as the indigenous khoe, and that slaves may have served a purpose other than as a source of unskilled labour, such as for artisan skills or for collateral.

Calumet Links, Johan Fourie and Erik Green

Economic History of Developing Regions