Domain: Rebel Civilian Violence and Accountability
The study of how rebel groups decide to target or spare civilians, with focus on how external sponsorship affects these incentives through principal-agent dynamics. Covers one-sided violence, atrocity commission, and civilian victimization by non-state armed groups.
Temporal scope: Post-Cold War period | Population: Rebel group-years with active conflict and potential external state sponsorship
Key Findings
- External state support (binary) for rebel groups significantly increases one-sided civilian violence (nbreg coeff on ‘support’, Table 1 Model 1, clustered on dyad), consistent with the moral hazard hypothesis: funded rebels reduce efforts to maintain civilian cooperation. (positive, strong)
- Democratic sponsors (democ_supdum) significantly reduce the positive effect of external support on civilian abuse (Table 1 Model 2), supporting the principal-agent moderation hypothesis: democratic principals impose human rights conditionality that constrains rebel agency slack. (negative, strong)
- Multiple state sponsors (num_supp) increase civilian abuse (Table 1 Model 4), consistent with the collective action problem among multiple principals: more sponsors reduce effective monitoring, each free-riding on others’ oversight, resulting in higher rebel one-sided violence. (positive, strong)
- Human rights organization secretariat presence in sponsor states (hrosecretariat_sum) significantly reduces rebel civilian abuse (Table 3 Models 1-2), extending the democratic accountability argument: principled lobbying by civil society actors moderates sponsor permissiveness toward rebel atrocities. (negative, strong)
- The interaction of proportion democratic sponsors × number of sponsors (perc_dem × num_sup, Table 1 Model 6) shows diminishing returns: as the number of sponsors grows, even a high proportion of democratic principals loses disciplining capacity, revealing the limits of democratic conditionality under collective action constraints. (conditional, moderate)