We present evidence that ethnic fragmentation explains variations in per capita income, institutions, and schooling better than income inequality when both are treated as endogenous. To do so, we identify instruments for ethnic fractionalization and income inequality based on historical experience. Using instrumental variables estimation, we find that ethnic fractionalization explains the level of income both when income inequality is included as a control in the estimation and when it is not. However, we find no evidence that income inequality affects the level of income when ethnic fractionalization is properly treated as an endogenous variable. We have similar findings when other development outcomes such as schooling or proxies for institutional quality are used as dependent variables. These results are robust to various controls and changes in the sample size and suggest that some of the previous findings regarding the effect of income inequality on development should be attributed to ethnic fractionalization.