Dangerous Territory: A subnational study on how territorial control affects violence against aid workers in Africa.

Abstract

Violence against aid workers has become a major concern for most humanitarian agencies. Adding to the destruction and chaos of conflict, deliberate attacks on humanitarian staff and facilities cause maximum damage to the health of populations. As long as they are being targeted, aid agencies cannot operate effectively and help the most vulnerable. This also poses a severe problem to the international community because deliberate attacks undermine the fundamental principles of humanitarian action. Irregular warfare, meaning governments fighting armed groups mostly in their own territory, has become the norm. These modern wars pose a challenge to the enforcement of international law and humanitarian principles. This study shows that micro-dynamics leading to attacks against aid workers are closely related to territorial control of armed groups. Government and rebel groups greatly benefit from the services provided by humanitarian agencies. Most armed groups are highly dependent on support from the civilian population. The delivery of crucial services to civilians makes aid agencies important actors when securing the support of civilians for whichever actor controls a territory. Renouncing them from aid deliveries may undermine popular support. This thesis argues that violence against aid workers can be a strategy to undermine the enemy's civilian base of support and is thus less likely in territory where an armed group has full control. In contested territory, however, in which no armed group has full control and compete over popular support, violence against aid workers is more likely. In order to study micro-dynamics in conflicts, a quantitative study on the subnational level in Africa is undertaken. The subnational level of the study is based on 0.5 longitude/latitude gid cells that rasterize the entire African continent. To ensure better comparability, the cells are pre-processed with coarsened exact matching. The results show that violence against aid workers is most likely in contested territory, as opposed to government- or rebel-held territory. The robustness prevails over a series of model specifications and robustness tests with alternative models. This analysis may help aid agencies to better plan their operations in conflict-affected zones.