This paper analyzes the relationship between ethnic fractionalization, polarization, and conflict. It argues that ethnic fractionalization, while related to economic growth, does not significantly explain the incidence of civil wars. Instead, ethnic polarization is a more relevant measure for understanding the likelihood of conflict. The paper develops a new index of ethnic polarization that satisfies the basic properties of polarization. Empirical results show that this index is a significant variable in explaining the incidence of civil wars, and the results are robust to alternative data sources and specifications. The paper also discusses the differences between ethnic fractionalization and polarization, showing that polarization better captures the potential for conflict. The results suggest that ethnic polarization is a more appropriate measure for understanding the relationship between ethnicity and conflict than ethnic fractionalization. The paper concludes that the lack of explanatory power of ethnic heterogeneity on civil wars is due to the use of an index of fractionalization rather than an index of polarization.This paper analyzes the relationship between ethnic fractionalization, polarization, and conflict. It argues that ethnic fractionalization, while related to economic growth, does not significantly explain the incidence of civil wars. Instead, ethnic polarization is a more relevant measure for understanding the likelihood of conflict. The paper develops a new index of ethnic polarization that satisfies the basic properties of polarization. Empirical results show that this index is a significant variable in explaining the incidence of civil wars, and the results are robust to alternative data sources and specifications. The paper also discusses the differences between ethnic fractionalization and polarization, showing that polarization better captures the potential for conflict. The results suggest that ethnic polarization is a more appropriate measure for understanding the relationship between ethnicity and conflict than ethnic fractionalization. The paper concludes that the lack of explanatory power of ethnic heterogeneity on civil wars is due to the use of an index of fractionalization rather than an index of polarization.