Investigating the ability of peacekeeping (PK) missions to reduce levels of violence against civilians (VAC) requires an assessment of whether these deployments control the spread of anti-civilian actions. Previous studies contend that PK operations prevent contagion and reduce the scope of rebel violence by reducing their external shelters and restricting their ability to move (Beardsley, 2010; Beardsley and Gleditsch, 2015). However, other research has drawn on case studies to demonstrate that PK missions might contribute to the relocation VAC into areas that peacekeepers cannot reach. Examples of this include the eastern part of the DR-Congo during the Ituri and Bukavu crisis in 2003-2004, and the trans-border area between Darfur (Sudan) and Chad (Clarik, 2008; Seybolt, 2012).
Drawing on these analyses, it is possible that deploying a PK mission will foster the relocation of anti-civilian activity, and that higher levels of VAC will subsequently be found outside of the mission’s scope. Analysing the locations of VAC, two opposing patterns are evident: in some cases, VAC spreads (as was the case in the Ituri crisis, and the Kivus, both in DR-Congo), while in others, VAC is relatively contained (Darfur, Sudan; Mali). Both these patterns are also conditioned by an overall increase (Darfur, Sudan; Ituri crisis, DR-Congo) or decrease (Mali; Kivus, DR-Congo) in VAC event levels. As such, even in cases of the geographic containment of VAC, this does not necessarily correspond to an overall reduction of levels of anti-civilian violence, nor does a spread imply an absolute increase in civilian targeting.
In the case of Darfur, there has been strong criticism of the role of the UNMIS (2005-2011), and the UNAMID (2007 – present) missions, and their lack of success in containing anti-civilian violence. A report from Human Rights Watch highlighted the inability of these missions to prevent cross-border violence, notably towards refugees in several Chadian camps (HRW, 2006). More recent analyses have contended that in Darfur, despite the deployment of UNAMID, VAC remains a persistent feature of conflict. Clarik (2008) has shown that violence progressively shifted towards the Chadian border from 2004 to 2008. Similarly, he demonstrated that levels of anti-civilian violence increased on the Chadian side of the border, particularly in 2008.
An analysis of ACLED data from this region confirms this pattern during UNMIS deployment and the first years of UNAMID (2007-2009), when there were significant levels of VAC near the border in Chad (see Figure 1). However, over time, there has been an evolution of the violence pattern away from the initial spill-over to Chad to subsequent geographic containment in Darfur. Whereas levels of VAC have increased overall in Darfur, this has coincided with a decrease in its geographic spread, and a concentration within Sudan. Since the end of 2010 and the withdrawal of MINUCAT – tasked with protecting refugees alongside Eufor Chad/CAR – not a single anti-civilian violence event has been recorded in the eastern parts of Chad. New events that occurred in 2013 onwards in CAR were not directly related to the Darfur crisis. This indicates a reignition of the Darfur Crisis in a different form than that which was experienced in earlier years.
Despite UNAMID’s inability to ensure the effective protection of civilians, anti-civilian violence has in recent years been geographically contained within Sudan. The extent to which this can be directly attributed to success on UNAMID’s part is unclear; it appears instead that the most important factor in shaping this geographic profile was the changing diplomatic relationship between Chad and Sudan. Attempts to normalize relations between Chad and Sudan in January 2010, contributed to the return of rebel groups in either country (Sudan Tribune, 2010). This agreement also allowed the deployment of 1,500 soldiers on a Joint Sudan-Chad border force to prevent rebels’ cross-border movements (International Crisis Group, 2010). While ending this proxy war by suppressing support to rebel groups, this agreement may have contributed to the concentration of anti-civilian actions within Darfur, where the violence was still on-going.
A similar pattern is evident in Mali, where the dynamics of violence support previous findings that VAC can be geographically contained following intervention. Although Tuareg rebels and Islamist groups were primarily responsible for anti-civilian violence in the northern regions of Gao, Kidal and Timbuktu, more limited civilian targeting occurred further south. Several anti-civilian attacks were also committed by government forces in Bamako, during the period immediately surrounding the March 2012 coup. However, by early 2013, enforcement deployments (French Operation Serval alongside AFISMA) contributed to a progressive containment of anti-civilian violence in northern regions. Challenges faced by the Tuareg and Islamist groups forced them to retreat further in the north.
As Figure 2 shows, three distinct stages of VAC can be identified in Mali. First, from January 2012 to November 2012, high VAC was recorded throughout the whole country. Second, the intervention (December 2012 – March 2013) of French Operation Serval forces (alongside AFISMA operation) contributed to the concentration of VAC in the north-central part of Mali, especially in Mopti and Segou regions, during a period when non-state armed groups still controlled several cities further north. Finally, when MINUSMA was deployed in April 2013, levels of VAC became concentrated in the extreme north in Gao, Kidal and Timbuktu.
In addition, neighbouring countries have not observed any significant spill-over of VAC due to displacement of regional armed groups such as AQIM. Mali therefore illustrates how the scope of VAC can be reduced and contained. Once more, the extent to which PK forces are responsible for this reduction in overall levels of VAC is debatable: the role of French and AFISMA troops appears to have been more important in reducing and containing the spread of violence, as well as the normalization of the political situation. MINUSMA’s primary contribution appears to have been sustaining that containment and preventing a re-diffusion.
In contrast to this first pattern, some countries have observed a diffusion of VAC during PK deployment. Research from Liberia show that VAC occurred in areas where peacekeepers were not present in the first stage of UNMIL deployment (Sept.-Oct. 2003), mainly in the eastern part of the country. Later, the absence of police forces can account for the recurrence of VAC in Grand Bassa and Montserrado, notably in May 2005 (Amnesty International, 2005). A study of DR-Congo further illustrates this dynamic of spreading violence. Following the April 2003 withdrawal of Ugandan MONUC peacekeepers from Ituri (Orientale Province, DR-Congo), the French-led Artemis mission was deployed in June 2003 in Bunia, to stabilize the situation. More than 500 civilians had been killed in early May 2003, and MONUC’s remaining forces were unable to prevent any of those massacres, revealing a vacuum of power (Holt et al., 2009). Following Artemis deployment in June, the situation in Bunia and the main towns in Ituri became more stable, but remained violent. MONUC forces redeployed from September 2003: out of 10,800 MONUC peacekeepers, 4,700 were deployed near Ituri and 4,000 in Maniema province. The remaining 2,100 were allocated to protect other areas in the Kivus (IRIN, 2004).
As Figure 3 shows, from January 2003 until August 2003, anti-civilian violence was mainly concentrated in Ituri (Orientale province) and in South Kivu. Progressive redeployment from September 2003 subsequently contributed to a relocation of VAC to the Kivus. Although fatality levels were 8 times lower (on average, 16 per week from January 2003 until August 2003, then 2 per week until June 2004), the geographic spread of VAC was wider. From September 2003 until June 2004, 75% of anti-civilian violence occurred in the Kivus, compared to 54% from January 2003 until August 2003. One possible explanation for this is that the redeployment of troops into Maniema and Oriental provinces instead of the Kivus, contributed to a shift in the security context in favour of rebel groups in the east. In the absence of the strong PK presence in the Kivus, non-state armed groups were more likely to engage in anti-civilian violence. MONUC’s ability to protect civilians in different areas appears to have been limited. At the end of May 2004, MONUC troops in Bukavu (North Kivu) struggled to limit the advance of local rebels groups, before scaling down their operations out of fear of jeopardizing the on-going peace transitional process (International Crisis Group, 2005). Several hundred civilians were forced to flee when rebel forces looted and burnt a market near Bukavu. Shifting the balance of power by concentrating PK forces within cities and specific areas appears to have facilitated a new civilian crisis.
A similar situation was observed in DR-Congo starting from March 2013 when the United Nations expanded MONUSCO’s mandate by deploying the Force Intervention Brigade to target specific armed groups (M23 and LRA among others). The anti-civilian actions committed by those armed groups spread within DR-Congo. With the exception of the LRA’s incursions and settlement in CAR, there were very few subsequent trans-border instances of VAC, which instead diffused primarily within DR-Congo itself. Levels remained extremely high in the Kivus, but anti-civilian violence progressively increased in neighbouring Oriental and Katanga provinces.
As Figure 4 shows, during MONUSCO’s robust mandate period (July 2010 until January 2013), North Kivu and South Kivu witnessed nearly 75% of the VAC carried out by these select active groups. By contrast, over the expanded mandate period (March 2013 – present), the Kivus witnessed only 41% of the anti-civilian violence. At the same time, levels were increasing in neighbouring provinces: Orientale witnessed 25.3% of these groups’ VAC in the expanded mandate period, compared to 19.5% in the robust phase. Similarly, Katanga witnessed 29.3% of these groups’ VAC during the expanded mandate period, compared to 5.6% during the robust phase (although levels in Katanga had already increased prior the expanded mandate). This may be explained by the strategic deployment of the PK forces, where peacekeepers were less numerous in Katanga – with 450 out of 22,000 peacekeepers in early 2014 (Reuters, 2014). Redeployment started mid-2014 (IRIN, 2014) leading to a net reduction in Katanga’s share of violence.
In closing, analysis of ACLED highlights two distinct patterns in the geographic location of anti-civilian violence during PK missions. In some cases, anti-civilian violence is primarily contained in its original territory. In Mali and Sudan, VAC became progressively concentrated in the northern territory and Darfur respectively, with limited evidence of widespread national or cross-border diffusion, although political and other factors coincided with PK deployment to contribute to these dynamics. A second pattern is observed in DR-Congo. The diffusion of VAC appears to take place when the balance of power between armed actors shifts and where areas have low levels of PK troops. This may contribute to targeted rebel groups facing opposition in strongholds (as during MONUSCO deployment) being incentivised to split into small units and spread into other areas fostering a VAC diffusion.
One reason for these dynamics includes a lack of troops (both soldiers and police officers) and their strategic concentration in specific areas, contributing to a vacuum of power elsewhere. Deploying troops and fighting volatile armed groups in places such as DR-Congo might therefore foster the diffusion of VAC. By contrast, in Mali, enforcement deployments led to the quick defeat of armed groups and, to a lesser extent, the stabilization of the political situation in large parts of the country, contributing to a significant decrease in VAC. These findings reflect the importance of micro-level approaches to understanding the success of PK mission in protecting civilians.
This report was originally featured in the September ACLED Conflict Trends Report.