Institutions and Geography: Comment on Acemoglu, Johnson and Robinson (2000)

This paper responds to findings by Acemoglu, Johnson and Robinson (2000) that suggest weak institutions, but not physical geography and correlates like disease burden, explain current variation in levels of economic development across former colonies. Using similar data and expanding the sample of countries analyzed, our regression analysis shows that both institutions and geographically-related variables such as malaria incidence or life expectancy at birth are strongly linked to gross national product per capita. We argue that the evidence presented in Acemoglu, Johnson and Robinson is likely limited by the inherently small sample of ex-colonies and the limited geographic dispersion of those countries.