Search

Many autocrats govern with an in-group, whose support must be secured, and an out-group, which is subject to repression. How do autocrats exploit in-group/out-group dynamics to secure their survival? One strategy, we argue, is to broadcast out-group repression to the in-group as a signal of the regime's capacity for violence. Empirically, we focus on China, where the government represses ethnic Uyghurs in Xinjiang. Drawing on 1 million articles from six propaganda newspapers, we show that the regime broadcasts out-group repression to urban elites on each anniversary of the 1989 Tiananmen Square massacre, when 10% of Beijing residents joined anti-regime protests. To understand its effects, we conducted a survey experiment balanced on the national census during the June 2020 Tiananmen anniversary. Using a list experiment to mitigate preference falsification, we show that CCP propaganda about Uyghurs during the Tiananmen anniversary discourages protests among politically engaged urban elites because they fear repression. [ABSTRACT FROM AUTHOR]

When do autocrats employ propaganda to threaten citizens with repression? Do threats of repression condition citizen behavior? This article develops a theory of propaganda-based threats in autocracies that builds on insights from experimental psychology. It argues that even credible threats of repression are costly, and so are reserved for moments when collective action is most likely. Since threats of repression are employed sparingly, the authors also expect them to be effective. The theory is tested using data from China, the world's most populous autocracy. The study analyzes all 164,707 articles published between 2009 and 2016 in the Workers' Daily, a state-run newspaper that focuses on domestic issues and targets a non-elite audience. It finds that the Chinese government employs propaganda-based threats of repression primarily around the anniversaries of ethnic separatist movements in Tibet and Xinjiang regions. Using an instrumental variables strategy, the study shows that these threats decrease protest rates by a substantively meaningful margin. [ABSTRACT FROM AUTHOR]