Rebel Finance and Sexual Violence

Rebel Finance and Sexual Violence Justin Conrad, James Igoe Walsh & Beth Elise Whitaker University of North Carolina at Charlotte November 14, 2015 (UNC Charlotte) PSSA 2015 1 / 12

Sexual Violence in Wartime Cross-national studies examine sexual violence during wartime and conclude that multiple causal mechanisms predict its occurrence (e.g., Wood 2006, Leiby 2007, Cohen 2013). Most studies conclude that, despite principal-agent problem, widespread sexual violence is frequently strategic in nature. Leadership has the authority to either encourage or restrain such violence. Variance in the magnitude of violence is driven by variance in incentives for leaders to either encourage or restrain. (UNC Charlotte) PSSA 2015 2 / 12

Sexual Violence in Wartime Organizations may deliberately encourage sexual violence to: 1 Coerce or punish locals, especially to prevent cooperation with adversary (e.g., Allen 1996, Brownmiller 1975, Seifert 1994, Kalyvas 2006). 2 Create social cohesion among group members (Cohen 2013). But widespread violence may stoke, rather than suppress, dissent. Under such circumstances, leadership has powerful incentives to curtail violence against civilians. (UNC Charlotte) PSSA 2015 3 / 12

Material Endowments and Rebel Violence Weinstein (2007) argues that groups with greater endowments are likely to engage in more violence against civilians because: These groups attract more opportunistic recruits who are less committed to the goals of the organization. These groups rely less on the civilian population for economic support. The consequences for violence against civilians is less severe for these groups. By contrast, groups with fewer endowments must cultivate a cooperative relationship with the local population. (UNC Charlotte) PSSA 2015 4 / 12

Extortion, Smuggling and Sexual Violence Natural resource endowments should not consistently influence the level of sexual violence. The method by which endowments are exploited should influence the relative degree to which groups rely on the civilian population. The degree to which groups rely on the civilian population determines the incentives for group leaders to restrain violence. (UNC Charlotte) PSSA 2015 5 / 12

Extortion, Smuggling and Sexual Violence Groups engaged in extortion of natural resource producers rely less on the local population. Extortion of existing production sites implies some level of control. Sites are likely to be geographically concentrated (Le Billon 2012). Such arrangements typically lead to long-term revenue streams (Sabates-Wheeler & Verwimp 2014). Groups engaged in smuggling of natural resources must rely heavily on local populations. Smuggling routes encompass much wider, diverse geographic territories. Includes more diverse populations and intermediaries. More sensitive to government counteroperations. Requires a great deal of discretion (Adler 2013). (UNC Charlotte) PSSA 2015 6 / 12

Hypotheses Hypothesis 1: Rebel groups that engage in extortion of natural resource production commit more frequent acts of sexual violence. Hypothesis 2: Rebel groups that engage in smuggling of natural resources commit less frequent acts of sexual violence. (UNC Charlotte) PSSA 2015 7 / 12

New Data on Rebel Finance Independent variables drawn from Natural Resources and Conflict (NRAC) Dataset. Unit of analysis: conflict dyad year. E.g., LRA-Uganda 2012 Temporal domain: 1990-2012 Key independent variables: Extortion: rebel group engaged in extortion of natural resource production. Smuggling: rebel group engaged in smuggling of natural resources. (UNC Charlotte) PSSA 2015 8 / 12

Research Design Dependent Variable: Prevalence of Sexual Violence Committed by Group (0-3 ordinal scale) State Department Human Rights Watch Controls Abduction Forced Recruitment Territorial Control Population Democracy Ln(GDP per capita) Estimation: ordinal logit; robust standard errors clustered on conflict dyad (UNC Charlotte) PSSA 2015 9 / 12

NRAC Data Table: Natural Resource Exploitation Strategies in NRAC Dataset Frequency Total Observations Percentage Any Strategy 625 1341 46.61% Extortion 423 1341 31.54% Smuggling 432 1341 32.21% (UNC Charlotte) PSSA 2015 10 / 12

Substantive Effects Figure: Substantive Effects on Sexual Violence 400 350 300 250 200 150 100 50 0-50 -100-150 Extortion Smuggling Abduction % Change in Odds of Higher Category of SV (UNC Charlotte) PSSA 2015 11 / 12

Results & Implications Groups that extort natural resources engage in higher magnitudes of sexual violence. Groups that smuggle natural resources engage in lower magnitudes of sexual violence. These effects are independent of the strength or size of the organization. These effects are independent of other causal mechanisms, such as whether the group uses abduction to staff its ranks. (UNC Charlotte) PSSA 2015 12 / 12