Search
Browse By Day
Browse By Time
Browse By Person
Browse By Mini-Conference
Browse By Division
Browse By Session or Event Type
Browse Sessions by Fields of Interest
Browse Papers by Fields of Interest
Search Tips
Conference
Location
About APSA
Personal Schedule
Change Preferences / Time Zone
Sign In
X (Twitter)
Heads of state signal their commitment to democratic or authoritarian values through public speeches. Such signals can constitute a breach of democratic norms (Schedler 2019) and can also indicate intent to undermine both formal and informal democratic institutions. By listening to what leaders say we can identify when their positions are out of line with existing institutions and therefore detect crises of democracy in the making. In this article, we develop an Illiberal Speech Index (ISI). Utilizing a machine learning approach, we train an LLM model on a corpus of speeches identified—through both expert coding and computational methods—as being stark examples of authoritarian speech. We then use that model to classify 25,025 speeches by 201 leaders in 74 countries between 1995 and 2022. The resulting scores place leaders and their speeches on a liberal-illiberal dimension. Time series analysis shows that the ISI can predict democratic decline and we test these findings robustness to country-level factors—such as existing democratic institutions—as well as international influences such as regional democracy. We show that public, authoritarian rhetoric provides an early warning signal of future autocratization.