Evaluating the Effects of International Criminal Court Prosecutions on Atrocities During Ongoing Armed Conflict

This dissertation examines the impact of International Criminal Court (ICC) prosecutions initiated during ongoing conflict on atrocities, with the goal of developing a typological theory specifying the conditions under which ICC prosecutions alternately prevent, exacerbate, or have no impact on wartime atrocities. This dissertation employs an inductive approach to theory building, using in-depth case studies of the impact of ICC action vis-à-vis two armed groups in the Democratic Republic of Congo (DRC) whose leaders were targeted for prosecution by the ICC, the CNDP/M23 and FDLR, to generate a set of testable hypotheses concerning the effects of ICC action. These case studies draw from over 100 original interviews with current and former members of these groups and other relevant actors in DRC, as well as statistical analysis of patterns of atrocities attributable to each group over time using data sourced from the Armed Conflict Location and Event Database. These case studies suggest variation in the impact of ICC action across stages of the legal process. In the CNDP/M23 case, I find that the issuance of an ICC arrest warrant for senior CNDP commander Bosco Ntaganda in April 2008 initially had negligible effects but subsequently contributed to an escalation in atrocities when Ntaganda, apparently fearing arrest after the conviction of his former commander, Thomas Lubanga, by the ICC and calls for his arrest by international donors and human rights organizations, reneged on a peace deal and instigated a new rebellion, the M23, in April 2013; however, Ntaganda’s surrender to the Court approximately one year later amidst internal tensions within M23 contributed to the prevention of atrocities by reducing M23’s capacity to perpetrate violence and contributing to its eventual military defeat. I find similar though not identical results in the case of the FDLR. First, I find that situation-level ICC action, encompassing preliminary phases before the issuance of indictments for individual leaders, had negligible effects vis-à-vis the FDLR. However, the execution of an indictment for senior FDLR political leader Callixte Mbarushimana initially exacerbated atrocities by further empowering radical leaders within the organization and provoking backlash against the ICC; in the long run, however, Mbarushimana’s arrest generated incapacitative effects that contributed to prevention. Finally, the issuance of an outstanding indictment for FDLR military commander Sylvestre Mudacumura had perverse effects, as it generated incentives for Mudacumura to spoil peace initiatives. I test the generalizability of these findings using statistical analysis of time-series cross-section data covering armed groups active in Africa from 2002 through 2010; this analysis yields three major findings, which are generally consistent with hypotheses derived from my case studies concerning the long-term effects of ICC action. First, this analysis indicates that situation-level action has negligible average effects on atrocities. Second, I find that outstanding indictments tend to exacerbate atrocities, and third, that the execution of ICC indictments contributes to the prevention of atrocities.