The use of violent coercion to repress unarmed protests, such as that seen during the Arab Spring, sometimes backfires on the government – an outcome called ‘political jiu-jitsu’. Examining unique global data covering extreme violence used by governments against unarmed protests from 1989 to 2011 (drawn from UCDP) and the Nonviolent and Violent Campaigns and Outcomes (NAVCO) data, this study aims to explain the conditions under which this outcome occurs. This study contributes to both the nonviolent action and one-sided violence literatures by further disaggregating this effect into both domestic and international outcomes, a distinction that has not previously been made in empirical studies. We find evidence that a pre-existing campaign infrastructure increases the likelihood of increased domestic mobilization and security defections after violent repression, but is unrelated to international backlash. Within ongoing NAVCO campaigns we find that parallel media institutions increase the likelihood of increased domestic mobilization and international repercussions after repression, and that this effect holds true for both traditional media and ‘new’ (i.e. internet-based) media. One of the novel contributions of this study is that we identify an important selection effect in the NAVCO data and the critical role of organizational infrastructure, especially communications infrastructure, in generating preference changes that create the conditions where killing unarmed civilians becomes costly for repressive governments. We conclude with a discussion of the potential implications of this study and avenues for future research.
