Esteban & Ray formalize the argument that conflict is likely to be more intense when individuals in a society are divided into two clearly identifiable groups where differences within groups are considerably smaller than differences between groups. They show that such polarization increases conflict, and they introduce a theoretical basis for its quantitative measurement. This article applies Esteban & Ray's (ER) measure of polarization to two international distributions: the world income distribution and the distribution of political systems. The article discusses extensions of the ER measure that render it independent of the number of groups in the system, and investigates empirically whether any of the three types of polarization are associated with conflict in the form of militarized interstate disputes. The results are rather inconclusive, despite the fact that some of the micro-level regularities consistent with the polarization and conflict argument have solid support in previous studies.