UN Troop Deployment and Preventing Violence Against Civilians in Darfur

ABSTRACT

Does the presence of UN peacekeeping force lower civilian fatalities at the local level? If it does, is it because of their coercive military capacity or for other reasons such as their roles in monitoring and reporting violent atrocities? To explore these questions, I study the deployment of peacekeeping units in Darfur and its impact on violence against civilians. Using original geocoded data of UN deployments before and after the intervention, I examine what aspects of such deployments impact one-sided civilian killings by government and rebel groups. Results indicate that deploying UN peacekeepers in an area restrains belligerent from targeting civilians. However, the military capacity of peacekeepers is not a significant predictor of violence against civilians. While their ability to defend themselves is extremely important for peacekeepers, these findings caution against the militarization trend in UN peacekeeping and seek to reshift focus on other substantive aspects of peacekeeping.

KEYWORDS:

• Peacekeeping
• UN
• Darfur
• responsibility to protect
• violence against civilians

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.

Notes

¹ In the words of former UN Secretary General Dag Hammarskjold, “Peacekeeping is not a job for soldiers, but only soldiers can do it.”

² Between deployment (treatment) and nondeployment (control) grid-cells.

³ Past studies discuss the role of major power in peacekeeping missions and the biases prevalent in UN Security Council Resolutions (Mullenbach 2005).

⁴ For instance, while many unarmed observers get deployed in war-prone areas to monitor the situation and provide an early warning, most armed peacekeeping units get deployed during the postconflict phase.

⁵ Rather than postaccord deployment, which is the focus of this study.

⁶ Critics argue that a large number of troops but with a low level of commitment can only exacerbate violent killings. For instance, Krain (2005) international indicates that when UN created safe haven sites in Bosnia-Herzegovina, but without a strong commitment to protect civilians, it only led to the mass killing of the Bosniacs.

⁷ That quality of troops in terms of their military effectiveness or coercive capacity does not deter armed actors from killing civilians.

⁸ The Government of Sudan, Sudan Liberation Movement/Army (SLM/A) and Justice and Equality Movement (JEM).

⁹ Signed by Chad, Sudan and Libyan government in February 2006.

¹⁰ A battalion has four to six companies, and has a total of 600–800 military personnel.

¹¹ The two main signatories were SLA (Minni Minawi) and the Government of Sudan. The major rebel forces in Darfur, namely the SLA/M (Abdul Wahid) and JEM did not sign the agreement.

¹² In Doha, where JEM became one of the signatories of the Agreement.

¹³ I also provide in the online supplementary file findings using UCDP GED data set, and the main results hold. I chose to use ACLED primarily because it provides a more comprehensive list of actors compared to UCDP GED, which is helpful in disaggregating rebel and government perpetrators.

¹⁴ Fatalities due to clashes or the number of peacekeepers that were killed by belligerents are not included in these dependent variables.

¹⁵ Table 1.1 in the online supplementary file shows the yearly frequency distribution of peacekeeping units (companies) in the region, from 2008 to 2015.

¹⁶ Once deployed, the peacekeeping units are stable until 2010. This is helpful in estimating the effect of the treatment in these deployed grid-cells. There is a change in only one of the grid-cells in 2010, which I exclude from the list of treated cells. But including this grid-cell does not have any significant impact on the main result.

¹⁷ Available in https://www.sipri.org/databases/milex (Accessed January 5, 2019). Toronto (2017) uses military expenditure per soldier from Correlates of War data set. But the COW data set extends only till 2012, and the correlation for each year between the two is around 85% for UNAMID troop contributing countries. Using military expenditure per soldier from the COW data set does not produce a substantively different result.

¹⁸ Kalma Camp, Otash camp and Al Salam camp are all in a grid cell in Nyala, Ed Daein camp, and Gereida, and Zamzam camp in a grid-cell near El Fasher. Data on IDP population is acquired from the Office of the United Nations High Commissioner for Human Rights.

¹⁹ AMIS presence in Darfur started trickling in since 2004. As of mid-2005, there were only 1647 protection force, mainly to provide security for 454 military observers as reported in the African Union Peace and Security Council report no. PSC/PR/2(XXVIII). It is not until the beginning of 2006, after the launch of AMIS II, that their strength is significant in various grid-cells. AMIS presence is therefore coded since 2006.

²⁰ See Wooldridge (2010:147–148) for discussion on how difference-in-difference technique for pooled cross-sections over two time periods that straddle across a policy change resemble a natural experiment, while also allowing for the option to add other covariates in the equation (Wooldridge 2010:151).

²¹ For further details on the equation, see Angrist and Pischke (2008:229).

²² The three dependent variables are: combined total fatalities by government and rebel groups; fatalities by government forces and militias; and fatalities by rebel groups.

²³ See online supplementary file for results with full models, and also for additional results with UCDP GED data set instead of ACLED, which also yields similar results.

²⁴ The third explanatory variable, the presence of peacekeepers, is not appropriate since this variable has a time-constant value of 1 for the number of years of deployment in grid-cells and a constant value of 0 for the rest. I, therefore, use only two other explanatory variables that time-varying and the Hausman test shows that fixed-effects model is appropriate for both.

²⁵ The data on IDP population (in ten thousand) comes from United Nations Office for the Coordination of Humanitarian Affairs in Darfur.

²⁶ Full models for Table 1.1, 1.2, and 1.3 are included in the supplementary material. Although models 1 in Table 1 includes variable AMIS, results in Table 1 will not be discussed since model specification for the table is mainly to measure the impact of UN peacekeeping units before and after their deployment.

²⁷ See also Sudan Tribune (2012) or Sudan Tribune (2011).

²⁸ Abrahms (2018: 74-81) discusses how rebel leaders of more successful groups are generally careful and strategic about causing harm to civilians.