Some recent empirical studies deny any direct effect of geography on development and conclude that institutions dominate all other potential determinants of development. An alternative view emphasizes that geographic factor such as disease ecology, as proxied by the prevalence of malaria, may have a large negative effect on income, independent of the quality of a country's institutions. For instance, pandemic malaria may create a large economic burden beyond medical costs and forgone earnings by affecting household behavior and such macroeconomic variables as international investment and trade. After controlling for institutional quality, malaria prevalence is found to cause quantitatively important negative effects on income. The robustness of this finding is checked by employing alternative instrumental variables, tests of over-identification restrictions, and tests of the validity of the point estimates and standard errors in the presence of weak instruments. The baseline findings appear to be robust to using alternative specifications, instrumentations, and samples. The reported estimates suggest that good institutions may be necessary but not sufficient for generating a persistent process of successful economic development.