С»ÆÊé

By: Jordan N. Hale

Abstract:Ìý

Ethnicity is a salient social identity in most democracies. Why is political rhetoric about ethnicity more common and more insulting in some elections than others? Comparative politics scholars often recommend less majoritarian electoral institutions for ethnically divided societies (Lijphart 1984; Horowitz 1985). They make this recommendation based on the cases of extremely divided societies where no ethnic majority exists, and studying macro level outcomes like the persistence of democracy and peace. Their analysis ignores the plurality of the world’s democracies, where an ethnic majority exists. They also did not have the technological ability or data necessary to compare rhetoric across countries and elections. Using current machine learning techniques on the text of political manifestos and tweets, I construct two datasets of ethnic rhetoric in 80 elections in 13 English and Spanish-speaking democracies, one of legislative candidate tweets, and the other of party manifestos. I find that the electoral institutions condition the effect of demographic diversity on the political salience of ethnic identity. Similarly, the impact of ethnicity’s salience on racist or xenophobic rhetoric is conditioned by the electoral system as well. Ethnic rhetoric predicts hate crimes and discrimination, and corresponds with less ethnically inclusive governing coalitions. However, there does seem to be an independent effect of electoral institutions on the inclusivity of governing coalitions apart from just the impact it has indirectly through rhetoric.