Analyzing nation-building policies toward non-core groups across Balkan states following World War I reveals strong support for my argument. Non-core groups supported by enemy powers and residing in states that favored the international status quo were most likely to be targeted with assimilationist policies. Exclusion was most likely for the non-core groups supported by external enemies, but which resided in states dominated by revisionist politics. Finally, non-core groups supported by allied powers were most likely to be accommodated.

Group size and the presence of a homeland did not significantly affect states’ nation-building strategies in the post–World War I Balkans. A difference in language and world religion between the non-core and core groups in a state made accommodation of the non-core group more likely than assimilation, but such differences do not help predict policy choices between exclusion and assimilation. The latter finding is rather surprising considering the well-established millet system and national historiographies that emphasize the role of religion in the nation-building process. Finally, whether a non-core group had been previously dominant does not seem to be a good predictor of nation-building policies. For instance, this finding implies that Christian states in the Balkans were not more likely to target Muslims, the previously dominant group of the Ottoman Empire, with exclusionary policies after World War I – when controlling for my argument. These findings qualify both status reversal and cultural distance arguments.