The Colonial Origins of Comparative Development: An Empirical Investigation: Reply

TLDR

Acemoglu, Johnson, and Robinson (2001) linked modern institutions to colonial mortality, a relationship that David Albouy contests as lacking robustness. Albouy weakens the original study by excluding Latin American and much of African data—about 60 % of the sample—and by using an inconsistently coded campaign dummy, whose correction undermines his conclusions. Our analysis demonstrates that controlling for outliers preserves the original results, rendering them robust even under Albouy’s extreme critiques. JEL codes: D02, E23, F54, I12, N40, O43, P14.

Abstract

Acemoglu, Johnson, and Robinson (2001) established that economic institutions today are correlated with expected mortality of European colonialists. David Albouy argues this relationship is not robust. He drops all data from Latin America and much of the data from Africa, making up almost 60 percent of our sample, despite much information on the mortality of Europeans in those places during the colonial period. He also includes a “campaign” dummy that is coded inconsistently; even modest corrections undermine his claims. We also show that limiting the effect of outliers strengthens our results, making them robust to even extreme versions of Albouy's critiques. (JEL D02, E23, F54, I12, N40, O43, P14)