On the Persistence of Persistence: Lessons from Long-term Trends in African Institutions

An influential strand of literature within economics and economic history called 'persistence studies' argues that low material living standards in African countries today were determined by institutional choices made in the past. However, the lack of consistent annual data on GDP per capita or institutional variables has meant that this literature has been largely silent as to whether their proposed relationships hold throughout the period it studies. This has made persistence studies vulnerable to criticisms of making leaps of faith or contributing to a 'compression of history'. Here, we draw on a data set of tax revenues for African polities for the period 1900-2015, with which we proxy the institutional capacity of a state. We then test whether some of the most influential determinants stressed in the persistence literature exert a consistent effect on our measure of institutions. Our findings suggest that the effect of population density and colonizer identity on institutions is not persistent. We find mixed results for precolonial centralization and ethnic fractionalization, while results for slave exports and settler mortality are more in accordance with theory. Overall, our results support the view that historical persistence should be measured, not simply assumed.