◆
THE HEART OF THE MATTER: SECURITY
Weak security institutions always carry risk of internal fragmentation and instability. South Sudan, with its history of divisions within the liberation movement, and of militia and proxy forces supported by Khartoum, was especially vulnerable. While it was not unexpected that a political crisis could lead to violence, its speed, scale and scope can best be explained from within the security forces, dating back to 2005.
After the CPA, the leadership of the SPLM/A decided not to change the names of the Movement and Army. Retaining the SPLM – just as other liberation movements had done elsewhere in Africa, would also help to keep public support for the party. The SPLM was a strong ‘brand’. Opinion polls confirmed this. Whatever their dissatisfaction on issues, the South Sudanese population’s support was rock solid. For the same reason, ‘SPLA’ retained credibility for the majority of people. During the interim period, few trusted that the war was over, least of all the SPLA-commanders, and it made sense to nurture the perception that the liberating force was still in full operation.
Come independence, however, I raised the issue of a name change with the leadership. A ‘South Sudan Armed Forces’, for example, would signal a professional, conventional force. I was told that this would come only after the transformation process had been completed. While this can be seen as a ‘chicken-and-egg’ issue, hesitation might simply have been fear of weakening the SPLA and its capacity to mobilize the cadres and people. The legacy of the liberation movement was anchored in the bones of every SPLA soldier and officer. They would not easily let go of their identity as freedom fighters. And when the border clashes with Sudan took place, they seemed to be back in their comfort zone; they would readily leave the irritating paperwork of the office and set out for the bush.
Disturbingly, this appeared to be the case also when the civil war broke out in December 2013. As we have seen, it started within the SPLA, at its two headquarters in Juba. The former Sudanese Armed Forces barracks south of the city, at Giyada, close to Nyakoron, is the main operations centre. The other, expansive headquarters, with most of the administrative departments and the Ministry of Defence is at Bilpam. A huge statue of John Garang looms above them, in his SPLA uniform, reminding every visitor that the SPLA remains a liberation army at the core. It was in the shadow of this image that leaders of the SPLA ended up fighting each other in a cycle of violence that literally spun out of control.
To understand what happened in the security forces, we need to look at the CPA-period, the failed integration efforts, lack of reforms, corruption and the crisis itself.
Ready to fight – again
After the CPA was signed, security was of highest priority. This was reflected in budgets, salaries as well as policies in relation to Khartoum. As late as in 2009, during a visit I paid to Juba, relations with the North were strained, and the issue of Abyei was particularly difficult. At a dinner Salva Kiir and three or four senior ministers made it clear that if the situation did not improve they would go back to the bush and fight. I was shocked. They did not need nice houses, cool cars and luxuries, they said; they had managed on nothing and could do so again. My attempts to remind them of the costs of war for their people and future generations failed miserably. In later discussions with members of the international community who believed the referendum or independence could be delayed, I recounted these and similar conversations. I knew that the alternative to enforcing the timetable in the CPA was not a better-prepared independence, but another war.
Budgets during the interim period reflected this. The overwhelming importance of external security justified the salaries of various forces. In 2006 a private was paid $125 per month. This was more than three times what his counterpart would have earned in the neighbouring Democratic Republic of the Congo. Officers of medium and high rank made between $2,000 and $4,000 a month. These munificent salaries – and politically important veterans’ pensions – explain why, from 2006 onwards, the SPLA alone accounted for over 40 per cent of the national budget. The overall security sector took up more than this, of course, but no one knows the true figures. After the austerity budget of 2009, official figures showed defence expenditure at 37 per cent of the total. Capital projects, including major procurements, were a major expense too, while relatively little went to operations and training. By April 2011, when privates’ pay had risen to about $220 a month,1 over 80 per cent of the defence budget was allocated to salaries.2 Ironically, soldiers’ pay was often in arrears, which contributed to instability in the ranks.
During the interim period it seemed to be assumed that military strength was measured by the number of soldiers, rather than by capacity and quality. A trained, well-equipped and mobile fighting force would likely have been much more effective in defending South Sudan where, however, as indeed elsewhere, reductions in personnel could be politically risky.
The CPA called for each side’s forces to withdraw from the other’s territory, and for establishment of Joint Integrated Units (JIUs). The latter had been championed by Garang and First Vice President Taha, but with Dr John dead and Taha sidelined in Khartoum, fully integrated units were never formed. Instead, the JIUs were to a large degree used for placement of forces of least priority for both armies; many of them former militia.
Indeed, as we have seen, ten days after Dr John’s death Salva Kiir took steps towards implementing his Big Tent strategy: offering militia leader Paulino Matip return to the SPLA under generous conditions.3 With the Juba Declaration of 2006, thousands of militia from multiple groups reintegrated with the SPLA. While Dr John also integrated some militia and probably would have seen the strategic wisdom in such a process, sources agree that he would not have done so on such a scale and in the same way:4 they claim that he would have been more concerned about the security risks and weakening the SPLA.5
Most militia leaders were from the Greater Upper Nile region, from Jonglei, Unity or Upper Nile State.6 One very credible source later reported that the number of actual fighters integrated in 2006 was wildly inflated. According to information obtained in Khartoum, the true number of militia was 15,000, not the reported 50,000. ‘We integrated shop-keepers, not soldiers’, he said.7 And while most militia leaders re-joined the SPLA, a few remained defiant;8 reintegration efforts continued through 2014.9 In numbers, Nuer predominated, and this affected the overall ethnic composition of the Army. These multiple reintegration processes depended on promotions in rank and economic incentives, nothing more. There was no reconciliation through dialogue, no settling of grievances.
Khartoum had always used cash and supplies of arms to buy loyalty and support from disgruntled local leaders and ethnic communities resentful of Dinka leadership of the SPLM/A. Now, it appeared, the South Sudanese had entered the market. Sudan’s former National Security Chief Salah Abdallah even complained in 2010 that southern militia had become so overpriced that Khartoum could not afford to compete!10 Left unsaid, but widely understood, was that leaders who had changed sides several times in the past were liable to do so again. The loyalty of these cadres to the SPLA – and to the President – was weak and opportunistic.
Reintegration of the militia had another worrying aspect. The intake of thousands of militia seemed to render the Army a social welfare system, taking care of old cadres and veterans. Edward Lino, a long- time SPLA commander, clearly frustrated by the 2013–14 crisis said that the SPLA had ‘never been a robust united force’ since it ‘started to incorporate militia’ in such ‘appalling numbers’; ‘each soldier was almost free … to choose’ his commanding officer: ‘there was nothing called “SPLA” It was divided and shredded into tribal formations adhering to individual commanders …’.11
While he may have exaggerated, the essence of Lino’s statement rings true for those who have closely monitored the poorly managed integration process of the SPLA.
After the Juba Declaration in 2006 a strong reform programme was formally adopted. The goal was a more professional national army, and a multi-ethnic Presidential Guard, trained together, streamlined and cohesive. The SPLA Commandos were formed as multi-ethnic units and trained to conduct special operations.12
Initial plans of creating a multi-ethnic Presidential Guard soon faltered. The President himself distrusted such an arrangement from a security perspective.13 In the end he, Riek Machar and Paulino Matip all had their own personal security details.14 Matip’s Nuer forces were never properly integrated.15 Even cabinet ministers did not want to relinquish their numerous handpicked bodyguards, and in effect a system of personal militias was accepted at the top.
For the SPLA as a whole, it was hoped that the example of the Commandos could be replicated. Plans were drawn up for amalgamating former militia into various units of the Army. But the process of absorption and reintegration across all units, training tens of thousands of soldiers, was so challenging that it soon met major capacity constraints. Senior officers were very concerned. In 2008 the SPLA leadership made serious efforts to stop additional reintegration; further dilution of the SPLA without progress in amalgamation and training would only serve Khartoum’s interests.16 ‘We were buying peace, it was a disaster to us,’ one officer said.17 But soon Salva Kiir and Riek Machar pushed for further integration of militia, lest Khartoum use those remaining outside the fold.18
From 2005 to 2011, adhering to impartiality between the two parties, the UN peacekeeping operation did not assist in this area. Programmes for demobilization, disarmament and reintegration (DDR) which were part of the UNMIS mandate, were not linked to such reform efforts, for example as an alternative to integration to the Army. Analysts have since pointed out how resource constraints prevented SPLA training plans from being implemented, and not least inhibited the successful integration of the other armed groups.19
At this time, support from key international players was also inadequate. Regional neighbours had the manpower, but not the resources, while it was the opposite with the donors. Only very late in 2007 and in 2008 did key donors start focusing on some types of SPLA reform programmes. But the window for defence reform and transformation is usually open for only a short time in post-conflict countries. And so it was in South Sudan. The motivation had been there at the outset, but the momentum was soon lost.
There were three important factors in the delay of reforms. One was relations with Khartoum. The need for a cohesive and effective fighting force in the event of another war had been a major incentive for early reform at the outset. By 2008, however, people in the leadership had started feeling that the CPA might be more sustainable than they had expected. One of the greatest incentives for SPLA reform therefore appeared to have been weakened. The second factor was the scale of the challenge. Multi-ethnic training and reorganization of units, as well as establishment of a professional command structure required more capacity than South Sudan could muster.20 The third factor was directly related to economic benefits. Budget allocations within the SPLA were not transparent, and allowed misappropriation of funds with impunity. Reforms would necessarily imply cleaning up the system; there were no offsetting financial incentives to do so.
‘Retail’ politics – corruption within the ranks
Corruption within the SPLA was not new. During the civil war there were no salaries, so cadres found other solutions, both for the Movement and for themselves. In such situations, food was like money, and could be traded for most other things, including arms. The World Food Programme needed, for example, one of the most extensive and advanced security operations in the humanitarian world to protect their convoys. Some comrades could acquire much wealth, and the top leadership of the Movement was at times characterized as a ‘military aristocracy’.21 Indeed, leading cadres themselves could in moments of lucidity admit that their generation was a ‘lost case’ when it came to management of financial resources; only a new generation of SPLM/A leaders would be able to set things straight.
After the signing of the CPA, opportunities for malfeasance were particularly lucrative as the military absorbed large numbers of armed elements, in the absence of proper registration processes. In 2006–7 the SPLA was estimated to employ 140,000 people, but the ‘core’ – soldiers – numbered only about 68,000.22 As independence approached, personnel numbers increased, with a spike during the last months to as many as 207,000.23
Credible sources have indicated that in December 2013 the SPLA’s payroll comprised approximately 230,000 soldiers. However, experts who did audits could not identify more than 170,000.24 At the same time, the government’s estimate for DDR also seemed unrealistically high.25 To date, the SPLA has not undergone a proper registration of all forces under its command. The SPLA’s own internal audit suggested a minimum of 40,000 ‘ghost soldiers’ whose salaries were either used to pay operational costs or pocketed by commanders.26
I discussed this in January 2012 with late Prime Minister Meles Zenawi of Ethiopia, with whom I had many conversations since the days of the CPA negotiations. Meles cautioned against going too far too fast: ‘The SPLA is not an army’, he said. ‘It has been there to keep people “salaried”.’ He worried that major clean-up efforts would backfire. It was a difficult dilemma. As Alex de Waal puts it:
Military commanders were both rewarding and defrauding their followers, by putting them on the payroll and cheating them of their full pay. For both patronage purposes and to lessen the dangers of the mobilization of the aggrieved, commanders assembled military units on tribal lines with the aim of maximizing personal loyalty. This is one reason why three attempts to institute a centralized roster of SPLA soldiers were thwarted.27
The opportunities for graft were also very significant. In the absence of proper procurement procedures, shady deals were easy to do. The international arms market is itself famously corrupt, and thrives on opaqueness.28 Both the seller and the buyer have an interest in secrecy.
It is rumoured (but not verified) that the first opportunity for big kickbacks came in 2006, and related to acquisition of Ukrainian T-72 tanks, a deal worth hundreds of millions of dollars.29 More Ukrainian tanks were purchased in 2007–8, and in 2010 several transport helicopters.30 Most of those interviewed aver that kickbacks were involved in all these deals, although a central source denies.31 A deal was also made in 2009–10 to procure 120 trucks at approximately $120,000 each. There are two versions of how many were delivered, 39 and 60. At least half of those paid for were therefore never delivered, the difference in cash having been syphoned away.32
In 2008 the Southern Sudan Government purchased air defence systems from Ukraine, a deal worth between $116 million and $199 million (the figures vary with informants).33 The price should have been much lower. According to credible accounts, about $71 million was deposited in Kampala and shared by three prominent members of the South Sudanese leadership. The air defence systems were never made operational, however, as some parts were not delivered, reportedly requiring an additional investment of another $40 million.34 When the border clashes with Sudan took place in 2012, other members of the leadership were furious that the SPLA had not acquired such deterrent capacity, despite the major investment that had supposedly been made several years before.
But the incentives to continue with the status quo were overwhelming. While reports and papers were produced, such as the White Paper on the SPLA in 2008 and the SPLA Act 2009, and decisions were made on critical issues such as enlistment age, service limits, retirement age, and pensions, there was little implementation. Regarding DDR almost nothing was achieved during the interim period.35
At the same time, defence spending continued to increase. In the 2013 audit of 2008 financials, only seven of 40 SPLA divisions reported payroll figures as required, even when repeatedly instructed to do so.36 Only in 2013 was the Auditor General given formal permission to review the overall accounts of the SPLA, but in early 2016, when this book went to print, nothing had happened.
From 2008–9 onwards, the Big Tent approach became increasingly controversial. There were tensions between senior commanders and former militia leaders who were inflating the officer corps and being promoted above them. Some of these latecomers had been in the service of Khartoum during the struggle, and it seemed unfair that they should do as well as or even better than comrades of long standing. Observers had started to refer to ‘rent-seeking rebellions’,37 by which commanders flip-flopped between Khartoum and the SPLA, trading their loyalty for a higher price in rank or funds. Their grievances might have been genuine, but so were suspicions about their manoeuvres, which called into question the whole Big Tent approach.
SPLA commanders blamed Khartoum for wanting ‘the SPLA to be diluted.’38 But the President consistently held the line on integration, convinced that this would consolidate peace and prevent Sudan from using militias to fight proxy wars against the SPLA and the South. Although there had been efforts at creating multi-ethnic units,39 some former militia commanders appeared to retain old networks that could be mobilized at short notice, or even kept former soldiers in close proximity. Primary loyalty continued to run through patronage or ethnic identity or clan and sub-clan affiliation. This led to an army fragmented and porous in structure and command, which could easily fracture along ethnic lines. Inflation of ranks was another factor; the SPLA reportedly had more than 700 generals.
Furthermore, the Big Tent approach left the whole integration process to the military. The political avenue was not used to address grievances, including ethnic concerns. As Edward Thomas says: ‘The SPLA was given the unmilitary task of providing a framework for national identity and national reconciliation.’40
Some SPLA soldiers granted leave to visit families ended up joining local units for convenience and local affiliation rather than returning to base. Some units lost many soldiers in this way, and therefore started in turn to recruit locally. This happened, for example, to the 4th Division, which was originally based in Eastern Equatoria, and composed mainly of Dinka and Equatorian soldiers and some militia. After it was transferred to Unity State, many soldiers drifted home, prompting local recruitment.41 The ambition to develop a competent multi-ethnic force in Unity, the critical oil-producing state, slowly gave way to a mostly Nuer force of uncertain quality.
While no one would share information about the ethnic composition of the SPLA, it is reasonable to assume that the perceived Dinka domination had declined and the proportion of Nuer risen significantly. Early in 2014 senior government officials claimed privately that former militia had constituted a majority of the Army by December 2013, and that that was at least one reason for the SPLA’s problems on the battlefield.42 Upon further analysis, that argument seems overblown. Crude but independent estimates indicated that approximately 40 per cent were likely to be Nuer, with the rest consisting of Dinka (probably about 35 per cent), Equatorians and others.43 Many of the Dinka and Equatorians were regarded as core SPLA, having been at the centre of the struggle throughout the civil war.
Whatever the case may be, the SPLA was less effective as a national army than it had been as a guerilla or liberation force. According to security experts, deterioration occurred primarily in 2006–7, when thousands of militia forces came in and efforts to integrate and train them properly faltered.44
International efforts – too little, too late
International support for reform came late, and focused on infrastructure and technical issues rather than on developing a multi-ethnic and cohesive force through systematic training at scale, revision of command structures, etc.45 A comprehensive strategic review early on would have helped, if supported by important donors and regional governments.46 While the UK and the US supported efforts to reform the SPLA, they did so late in the day. Pressure for a SPLA Transformation and Research Directorate established in 2010 was one such intervention. But it did not work effectively. External experts appeared not to understand the SPLA’s internal dynamics well, many cadres preferring a different approach.47 As we have seen, the inability of the European Union, including the UK, to provide direct support to military training because of sanctions against Sudan was one of the reasons.
After independence more efforts were made, and UNMISS tried to get an overview of the support provided to the SPLA now that it was a national army.48 The Mission was not mandated to play a role in this regard, and the US and the UK continued as main contributors. We were, however, mandated to support security-sector reform and had the main responsibility for assistance in DDR. Coherence among all stakeholders and coordination with other security-sector interventions were all the more important after independence for security sector reform to succeed.49 But in practice the exchange of information was limited, which was symptomatic of hesitation to coordinate. Regarding support from other sources, and in particular from the region, no government wanted to share details. There was no systematic transparency or complementarity of effort.
An ambitious programme was drawn up under the leadership of the DDR Commission and with support from us and others, but without real interest within the SPLA leadership. The programme was anyway too costly and rather unrealistic, and not adequately adapted to South Sudanese conditions.50 No consideration had been given to alternative approaches such as a pension fund and resettlement. Full registration of the SPLA was still on hold; transformation plans were adopted, but not implemented.51 By the time of the crisis, in late 2013, I had started discussions with security ministers on how these plans, including a total revamping of the DDR programme, could be moved forward. We were also discussing how security-sector coordination could be improved.
There was no high-level commitment. The leaders of both Government and Army seemed to prefer a more comfortable (and lucrative?) route, and hesitated over politically risky decisions. A country’s security situation is also often an impediment to reform, whether external or internal.52 Only in the absence of violent conflict can reforms, such as increasing civilian control of the military and implementing DDR, be pursued successfully.53
That the war with Sudan might re-erupt at any time could have reinforced the urgency for transformation of the SPLA to a professional national army. But the security threats and tensions were immediate and now had the opposite effect. They were a disincentive to reform and DDR. Experience immediately following independence, with military overflights, bombings and ground incursions provided a shaky basis for reform efforts, let alone for full transformation. Violence in Jonglei and militia activity in the border areas also complicated such efforts.
Other security forces – rule by gun, and not by law
After the signing of the CPA, the Southern Sudan Police Service (operating under different names after independence), the Prison Service, the Wildlife Service and Fire Brigade were also established. In their development many similarities with the SPLA were evident.
John Garang wanted all the cadres of the SPLM/A to have some kind of job after the peace agreement had been signed. In 2004, as Norway’s minister of international development, I brought in my friend Ashraf Ghani, who had then completed his tenure as finance minister of Afghanistan, to advise Garang and the leadership on critical transition issues. Ghani was adamant that peace could best be consolidated by ensuring that those who had participated in the struggle got some kind of pay or pension, at least for a transition period.54 This advice backed Garang’s strategy. The decision was soon made to provide all the old cadres, including veterans, with a salary by establishing various uniformed forces. It was a way of buying peace.
There was a clear hierarchy among the new services. To put it crudely, the perception was that first-class soldiers remained in the SPLA, second-class went to the police, and the rest ended up in the Prison Service, the Wildlife Service and the fire brigades. Transfer was not a matter of choice, and many soldiers found themselves demoted and unhappy. A number of militia were also integrated into the police after the Juba Declaration of 2006; several thousand police were transferred from the North and similarly integrated. Disabled veterans were formally assigned to the various services.
A clear illustration of this tiered system was the difference in salaries. By 2012, the disparity in salaries at all levels was significant, with SPLA colonels paid three or four times more than the National Police of comparable rank.55 Other terms of service were different, too.56 The Army carried much more prestige, both within the Movement and among the population at large. This led to disaffection in other security services.
Other security forces – same challenges, same ‘diseases’
Development of a competent police service was fundamental for Southern Sudan, a cornerstone for the rule of law in the semi-autonomous region during the interim period. In the territories under its control during the liberation struggle the SPLA had been police, investigator, judge and prison warden, with no independent oversight or checks and balances; the local commander was everything. Establishing a judiciary and police independent from the military was a major step.
But the nascent police service was incapable of dealing with internal security challenges, and the SPLA was forced to assume responsibility on many occasions. Internal security and protection of civilians and their property was not the military’s forte, with rather bad and heavy-handed management of security challenges and of inter-communal violence. But police primacy could be affirmed only when the capacity of the police service had been adequately developed.
Having very limited education, if any, and no training, new recruits to the police were also poorly motivated. Senior police officers told me that many recruits still saw themselves as soldiers, ‘on loan’ and awaiting return to the SPLA. This feeling was even stronger among former militia.57 That most security services still operated with military ranks, titles and hierarchy did not improve matters. Their uniforms, while different from the SPLA, were still military in design. This added to the feeling that they really belonged in the Army, and that their current situation was temporary.
With this background it is unsurprising that bad habits of the SPLA showed up, and even multiplied, in the other services. Not only was there less command and control and less discipline, but police personnel were also frequently accused of serious misconduct, drunkenness, unnecessary force, assault, bribery and extortion, all, of course, in an environment lacking transparency and accountability.
A case in point was a shooting incident in the troubled Tri-State area of Unity, Lakes and Warrap on 1 February 2012. After communal fighting had resulted in 78 deaths, a meeting of local authorities, state police leaders and relevant security services of Unity State took place at the Mayendit County authority’s compound, to prevent revenge attacks and a cycle of violence. UNMISS’s State Coordinator for Unity, several United Nations Police (UNPOL) and other staff were also present.
After the meeting started, four pick-ups arrived with armed men who started shooting indiscriminately. Uniformed personnel joined in. There were later rumours that soldiers from the SPLA’s 6th Division had joined the fighting, which continued for two hours. An estimated 50 people were killed, including many civilians. While some saw this as a revenge attack, others discerned ethnically motivated violence within the forces. UNMISS staff were caught in the crossfire, and one UNPOL officer was wounded. Only after several hours was one of our helicopters able to land to rescue them. Government officials told me that the whole thing was a misunderstanding: one side thought the other was attacking them, when no one was. Be that as it may, the incident certainly illustrated problems of command and control in the uniformed services.
While reform plans were written up for the security services, in particular the police and prisons, resistance from within was significant. Transformation of the Southern Sudan Police Service (SSPS) would ultimately involve at a minimum compiling accurate numbers and adjusting the size of the force.58 The same was true of the Prison and Wildlife services. While these services offered less scope for manipulating procurement contracts, there were ample opportunities for corruption in relation to payroll.
Critical measures were therefore delayed for years. Former UNMIS focused more on monitoring and less on capacity building. There was not much donor engagement, although the UK and US provided support. The SSPS was still embryonic, and could have been steered in the right direction. Observers regard this as a missed opportunity.59
It was only ahead of the elections in 2010 that a major reform effort was made. While UNPOL estimated the total number of police at 33,000 – but at most 60,000 as ‘useful and deployable’60 – efforts were made to train new recruits at Rajaf. As many as 6,000 recruits – men and women – were sent for training. Unfortunately, living and training conditions were bad, and accusations of human-rights violations soon ensued, forcing the UN Mission and donors to withdraw support.61
Recruitment continued to the SSPS and increased throughout 2011. Transformation of the SSPS to the South Sudan National Police Service (SSNPS) was imperative.
Transformation of the South Sudan Police
In late 2011, after independence, the South Sudan National Police Service was assumed to have approximately 50,000 personnel, although we knew of many ‘ghost’ names on the lists. Some personnel even kept two uniforms, and claimed salaries from both the police and the Army.62 Many officers in charge in rural areas were former militia who had been integrated into the SPLA after the Juba Declaration of 2006.
It was also only now, under the new minister of the interior, Alison Magaya, supported by UNMISS Police Commissioner Fred Yiga, that the registration process began. They made interesting discoveries. One police officer turned out to be a bull. A grandmother in her 70s had been listed as a captain – and a man; she claimed entitlement to the salary of her long deceased son. Colonels who were 12 years old were discovered, too – people had just added their children to the lists.63 The registration process was completed under the new Inspector General of Police Pieng Deng, who was appointed in January 2013 and was very supportive in this final phase. In an excellent example of what was possible, UNMISS had assisted with correcting the payroll, in cooperation with UNDP. The SSNPS had in fact 35,633 personnel. More than 16,000 ‘ghosts’ were identified, a number that reportedly translated into tens of millions of dollars, money that was now reallocated for equipment and infrastructure.
Getting the payroll right was only one important step. Now, a major transformation process was needed. The new inspector-general of police requested the support of UNMISS and UNPOL to assess the state of the police and provide data to help accelerate the transformation process. The UK’s Safety and Access to Justice Programme, in collaboration with UNPOL and other partners, had already led important strategic reforms, and these combined efforts ultimately helped to change mindsets.
On the basis of the newly registered personnel, the joint assessment included an analysis of literacy and language capacities and competence levels, disaggregated by rank, and identified five areas critical to the transformation of the SSNPS, including training gaps, capacity inadequacies, and institutional deficiencies.64 (When police cannot read the Penal Code, or check procedures for arrest, major problems ensue.) The assessment, made public in early 2013, revealed that only one-third of the police were literate in English, and one-sixth in Arabic. The rest were illiterate in all languages. In Unity and Jonglei as many as nine of ten officers were illiterate. And there had been a huge inflation in rank: the ratio of non-commissioned officers to constables was almost 1:1.
The assessment identified major challenges and recommended reforming the whole SSNPS, its legislative framework and operational systems (command and control, accountability standards, a standardized training system, etc.). The Service adopted all the recommendations, as did the new interior minister, Aleu Ateny Aleu, in August 2013. Soon thereafter, a Training of Trainers course began for 300 university-educated police personnel. The process of streamlining command-and-control structures commenced. In early October, all ten governors and the cabinet were briefed and approved a roadmap for transforming the SSNPS.
The director of the Prison Service approached UNMISS for similar support. During 2012–13, in line with the Mission’s mandate regarding the rule of law, we drew up strategic plans for the Service that for the first time included significant reforms. Early on, the minister of wildlife conservation and tourism, Gabriel Chansong Chang, had also requested UNMISS help in registration and cleaning up the payroll, but we had to complete the Police process first because of limited resources. The minister went ahead anyway, and was under serious threat of violence several times from his own staff and officers, illustrating the challenges such transformation processes could entail.
Reform of the entire National Security Policy in 2013 showed considerable promise. Contrary to prior efforts, donor countries, their experts, UNMISS, and advisers from the African region worked together in advising the government, under the leadership of the minister of national security, Oyay Deng Ajak. Through extensive consultations in all ten states and with civil society and international stakeholders, a truly inclusive process was completed, across party lines, and facilitated by the Mission. The draft National Security Policy that resulted was locally owned and, in its last draft, seen as one of the most advanced in Africa.65 It would ensure coherence across the various sectors, under the overall supervision of a National Security Council.
Before the crisis broke out, the policy was about to go to the President and cabinet, but it was stalled thereafter, presumably by the new minister of national security, Obote Mamur Mete, and the national security directors, who appeared to prefer a different approach. National Security legislation was instead presented for parliamentary approval but was, as we shall see, almost totally different from the pre-crisis draft.
With the violence in Juba in December 2013, the window of opportunity for reforming the SSNPS from top to bottom was closed, and challenges in the security sector increased exponentially. All security services subsequently divided along ethnic lines. The institutions were too weak; command-and-control lines too porous; and ethnic identity too strong for any of the security services to resist the tidal wave that now rolled over them.
Playing with fire – other developments prior to the crisis
In the absence of a robust army and transformation of the other security forces, the situation was very fragile. With a significant part of the Army comprising poorly integrated Nuer, the loyalty of the SPLA to its commander-in-chief, the President, a Dinka, would be in doubt in an escalating crisis. For the President, this was a risky situation. For others, this might have been seen as an opportunity.
In 2012–13 other factors came into play. After the events in Heglig in 2012, when clashes on the border prompted South Sudan to occupy this oil-producing area, the government called for a national mobilization of able men. Each state recruited 5,000 soldiers. Governor Paul Malong of Northern Bahr El Ghazal recruited more than his share. After the mobilization was cancelled, he wanted to retain the recruits in a training camp at Pantiit in Aweil North. Numbering as many as 14,000, they were all young Dinka and came primarily from the Governor’s state and neighbouring Warrap, the President’s home state.66 They were called ‘Mathiang Anyor’, a Dinka name for predator cat.67 The force was not funded by the SPLA, since it was not part of the Army, but was supplied by the Governor and the Bahr el Gahzal community with their own resources.68
On 14 February 2013 Salva Kiir issued a decree placing 117 senior officers on the paid reserve list. These included two lieutenant generals, two major generals, and over 100 brigadiers.69 Some had been serving in the civil administration; in fact 8 of 10 elected governors were ex-military.70 This, on top of the removal the previous month of some 35 senior officers, constituted the biggest shake-up in the Army leadership since independence. Those in civilian positions remained in office. Although seemingly a dramatic move, it had been long in the making. According to the SPLA’s highest officials, the relegations were the work of a committee, not ‘political’ in any way, but part of the reform effort.71
When it was proposed that a number of new recruits from Pantiit should reinforce the Presidential Guards in Juba, some senior officers in the SPLA and Ministry of Defence preferred reinforcement through internal redeployment. They suspected a hidden agenda.72 Paulino Matip’s forces, which had been officially under the auspices of the Presidential Guards, but had largely remained separate under his control, were divided after his death in 2012. Some were reportedly integrated into the SPLA proper, while others remained in the Presidential Guards. According to most informants, this meant that Nuer were now a majority in the Guards.73
Paul Malong and President Kiir appear to have agreed that a number of ‘Mathiang Anyor’ would be transferred to Luri, the SPLA training centre outside Juba, and become Presidential Guards. This was not the first time the Governor had stepped in to enhance the security of the President. In 2005–6, when Kiir had been concerned about plans for multi-ethnic presidential security, which included taking on a number of Garang’s bodyguards, Paul Malong and General Salva Mathok, Kiir’s relative and deputy chief of staff in the SPLA, had stepped in and recruited closer to home.74 Now, as danger threatened, Malong appeared to have taken action again. Tensions with Riek Machar were increasing, Malong worried that Nuer would dominate the Presidential Guards, and he wanted to restore balance with a large number of new recruits. When these were transferred to Luri, outside normal SPLA deployment plans,75 one senior security official said, ‘We smelled a rat.’76
‘Dot Bany’ – rescue the leader
While the first batch of recruits sent from Pantiit in Northern Bahr el Gahzal to Luri was a battalion of 800, senior officials at the time said that the total was three battalions.77 A smaller number appeared to be trained to become Presidential Guards,78 with others to be assigned to different duties in the Army. Major General Marial Ciennoung, the commander of the Presidental Guards, told another senior official that those graduated to join the Tigers later in 2013 were 700, almost a battalion.79 A number of interviewees claimed that the number was higher; others lower. Whatever the case may be, the special recruitment and training arrangement for Luri, as well as ethnically based loyalties, probably combined to strengthen the protective forces around the President, whatever duties they were assigned.80
The recruitment for the Presidential Guards at Luri led to tensions with the Army leadership and senior officials in the Ministry of Defence, strains that increased further when funding from the SPLA also was requested.81 It is not clear how the training of the recruits at Luri was funded. Some say the President had a separate security fund for miscellaneous purposes,82 or through fungible off-budget funds of National Security; others claim that fundraising efforts among the Bahr el Ghazal community helped; and others again that funds were diverted from the SPLA 5th and 8th Divisions, leading to delayed salary-payments.83 It might be that money came from all these sources.
Recruiting and training extra forces for the President’s protection was probably seen as an additional security measure at a time when tensions within the SPLM were rising. Prior to the rescinding of the Vice President’s delegated powers, Paul Malong was quoted as saying that he had sufficient forces to prevent Riek from ever leading the country. If Riek did not relent, Malong said, ‘Mark my words: I will push Riek and the Nuer to the other side of the river.’84 Whether this was just ‘tough talking’, is not easy to say. There were no other witnesses, which has made corroboration difficult. However, on two later occasions Malong made similar comments to others, increasing the likelihood that the information might be credible.
After Riek’s statement on 4 July in the Guardian, that ‘he was ready for a fight’, things only got worse. Some thought he was preparing militarily. A very senior figure immediately confronted him.85 Riek responded that by ‘fight’ he had meant only a political showdown, not a resort to arms. Despite encouragement to publish a clarification, however, he let it go. A number of people around the President seemed to take Riek at his (literal) word.86 The many Nuer in the Presidential Guards now only added to concern.
The forces trained at Luri to become Presidential Guards got special Dinka names, used interchangeably: ‘Dot Bany’, meaning ‘rescue the leader’ and ‘Gel Bany’ (‘protect the leader’).87 Sources say that a close confidant of Salva Kiir, former Chief Justice Ambrose Ring Thik, gave the Luri forces these names, either when they were brought to Luri,88 or on 16 December, as the crisis erupted.89 In 2012–13, and together with a few community leaders, he had reconvened the so-called Dinka Council of Elders, which in the 1990s in London had served as a focal point for resistance to John Garang. Now consisting mainly of well-connected people from Warrap, Kiir’s home state, they resented the political threat posed by Riek Machar.90 They often met in the President’s house.91 Sources are clear that the group planned a strictly political confrontation with Riek.92 The question remains whether they had links to others from the Bahr el Ghazal network in the security apparatus discussing what to do if things got out of hand – and whether such discussions included mobilization of the ‘Luri boys’.
Having picked up rumours from various sources, I raised questions with the President and ministers. We were told that although the first recruits were from the Bahr el Ghazal, the expanded Presidential Guard would be multi-ethnic. While there was arguably no need for expansion, most African presidents tend to have extensive personal security, so the repeated explanation was plausible. It proved not to be true. I followed up with security ministers, and got the same response. We should have pressed further. Towards the end of 2013 Paul Malong was spending more and more time in Juba, away from his home state.
The Presidential Guards were easily identifiable from their tiger-striped uniforms. The Imatong Tigers were the unit Salva Kiir first commanded in 1983. As we have seen, his code name during the struggle was ‘Tiger’, hence the name of the Presidential Guard and the choice of uniform for the press conference on 16 December. The Luri boys, or ‘Dot Bany’ were provided with different uniforms, also easily recognizable, but were still under the Tiger Battalion’s command.
The Luri boys’ arms and ammunition appeared to have come from contractors other than those supplying the SPLA, and most likely arranged through National Security. Their newest weapons may have come from China or Israel.93 The arrival of weapons from China led to speculation that something had been planned. However, these imports had been ordered long before the 2013 crisis.94
Most ‘Gel Bany’ spoke only Dinka. They were active during the fighting in Juba on 16–20 December, and easy to discern from their distinct uniforms and insignia.95 Both Tiger and Luri forces have been identified as perpetrators during a number of ‘sweeping operations’, house-to-house searches, targeted killings and atrocities against civilians.96 While some other units of the security apparatus tried to contain the situation then, others went on the rampage. Numerous eyewitnesses noted that many soldiers involved in the atrocities knew only the Dinka language. A well-placed source in the SPLA claims that Chief of General Staff James Hoth Mai was sidelined at this time, and that Paul Malong commanded units directly, both during the crisis in Juba and later. This involvement included circumventing normal command lines, directly commanding lower ranks and officers in the field.97 This has not been fully corroborated.
A number of credible sources with intimate knowledge of the events before and during the crisis avow that the SPLA leadership was not involved in any way. It appeared, however, that a network of old and retired commanders outside the formal SPLA command structure, all Dinkas originating from Northern Bahr el Ghazal and Warrap, immediately turned up and started giving orders.98 Four names were highlighted to me by very credible sources early on; later firmly corroborated through a number of interviews.99 These generals are believed to be in the inner circles of power. All of them had been put on the reserve list or relieved of their rank earlier in 2013. The AU Commission of Inquiry quotes Defence Minister Kuol Manyang referring to ‘a shadowy “group [that had] organized itself as Rescue the President. It killed most people here [in Juba] – from 15th to 18th. It was even more powerful than organized forces.”’100
It seems clear that the SPLA senior command was not involved. James Hoth Mai, SPLA Chief of General Staff and a Nuer, was immediately engaged in trying to contain the violence and regain control in the capital and other affected areas.101 He worked closely with Paul Malong, who had already relocated to Juba and SPLA headquarters, where he remained throughout the crisis. Some of those alleged to be involved in directing the most violent operations during 16–18 December in Juba had reputations for brutality.102 They were also regarded as close to the President. The above-mentioned four had already (on 20 December), three days into the crisis, been reinstated in their former military ranks along with some Nuer commanders, presumably so the latter would not defect.103 That this happened can be seen as public acknowledgement of their performance during the crisis.
While there have been many accusations that the killings in Juba resulted from elaborate advance planning of almost genocidal intent, supported by the fact that Juba had been divided into sectors,104 this has been impossible to substantiate. The capital had earlier been divided into sectors by the security agencies and forces in response to rising crime rates. The question is rather whether killings and atrocities were planned with the use of Dinka forces, whether from Luri, the Tiger Battalion, or the SPLA, all combined, or whether such forces got totally out of control.
On 12 December, three days before the crisis erupted, Oyay Deng, Pieng Deng and Taban Deng were invited to dinner at Governor Paul Malong’s house in Juba. Conversation centred on the political crisis. Paul Malong did not want Riek Machar to contest the chairmanship at the SPLM Convention (which was still expected in April or May 2014). He wanted Salva Kiir to stand in the 2015 election, and reportedly said that Riek could take over in 2020. If Riek refused, there would be fighting, he stated. The discussion was animated.105 On the 15th Malong went to urgently see Oyay Deng before noon, saying that he should abandon the others and join the President and that the division in the Movement would last a long time.106 Oyay rejected the offer. Paul Malong appeared to be aware that something was going to happen.
At around 7 that evening I met one of the security ministers at his house. I was very concerned about the developments at the National Liberation Council, and wanted him to talk to the President about trying to reconcile dissidents. He seemed amenable. In my presence, the minister got a call from security headquarters; he was needed urgently. He later told me that the security organs of the government knew that something was going to happen that evening.
The findings of the subsequent UNMISS human rights investigations suggested ‘coordination and planning’ had taken place in relation to some violent incidents.107 According to the African Union Commission of Inquiry:
The evidence thus suggests that these crimes were committed pursuant to or in furtherance of State policy [...] [proving] […] the ‘widespread or systematic nature’ of the attacks. The evidence also shows that it was an organized military operation that could not have been successful without concerted efforts from various actors in the military and government circles.108
Did civilian leaders and military commanders instruct forces to carry out attacks on Nuer civilians? This can only be clarified through independent criminal investigations.109
Political rallying turns violent
What was the situation on Riek Machar’s side? Was there planning, as some have alleged, for a military coup? There had long been fear that Riek would raise a rebellion, and there was unease whenever he ventured to areas in the Upper Nile region. UNMISS was confronted several times with rumours that we were going to transport him to certain places in one of our helicopters. But we followed UN rules, which stipulated that only travel that could be justified by the mandate of the Mission could be facilitated; Riek did make requests for transportation after leaving office, but since these were usually not justifiable they were rejected. That this matter went all the way to the President, and that he personally confronted UNMISS leadership, showed how tense the situation was.
After Riek Machar and his colleagues were dismissed, stories had begun circulating in Juba about the possible acquisition of arms and equipment by his followers. The allegation was that Taban Deng Gai had started operating when he returned from Dubai, about a month after he was ousted as Governor of Unity State, that is, in June 2013,110 before Riek Machar was dismissed. Such activities supposedly increased from September 2013.
There was nothing new in rumours like this. Machar had already been accused at least eight times of planning a coup, Majak D’Agoot six times, Oyay Deng four, etc.111 In July 2012, when Riek Machar, Oyay Deng, James Hoth Mai and Majak D’Agoot were all at the airport to bid the President farewell on one of his trips, a security officer became suspicious when they huddled in a car. Their desire for air-conditioning was soon inflated into coup plotting. Security forces were put on high alert, and checkpoints were established all over the capital. It took almost two weeks to calm things down.
South Sudanese have an incredible ability to spin rumours, and not least, conspiracy theories. As with their Sudanese brothers in the north, the security and intelligence apparatus are among the worst at this. Competition between services results in a race to be first to the top with the latest – often unverified – information, thus feeding a sense of paranoia. I experienced many examples of this unfortunate tendency, leading at times to faulty decisions.
To what extent this played out in the case of Riek Machar is difficult to tell. His movements and conversations were monitored. A lot of his activities seemed suspicious to security officers and government officials, including his speeches to the Nuer community, and two senior officials cautioned him in October and November 2013.112 SPLA commanders of Nuer origin were often observed coming to see him. Machar himself claims that all such meetings were purely political, and had nothing to do with plots or suspicious activity.
As early as June, in Dubai, after having been dismissed as Governor, Taban Deng Gai had told a comrade in reference to the President that they ‘must do something’.113 Several people heard him talking this way.114 One credible source puts Taban Deng in Riek’s house very often during this period, and ‘pushing very hard’; Riek in turn was overheard several times saying that he did not want a repeat of 1991,115 and that Taban Deng was a ‘loose cannon’.116
Several sources in Khartoum, including within the security apparatus, and others within the SPLM leadership refer to contacts with Taban before the crisis. Apparently, he was asking for military support through a business colleague with close links to military intelligence in Khartoum.117 I was unable to interview Taban about this. Interestingly, Khartoum rejected these requests, at least at the time. Most interviewees confirm that Taban was mobilizing. The extent to which he was trying to start a rebellion through acquiring weapons is disputed, however.118 Such attempts could also be part of an effort to rally support in the event that things deteriorated further and a confrontation was unavoidable. In other words, these are not necessarily evidence of a plan to oust a President from power, but may have been preparations in case of conflict.
It is interesting that friends and foes alike – excepting members of the current government – agree that Riek would not try to orchestrate a military coup, not least because of his limited following in the military and even among the Nuer. But a number of them are less certain about Taban. One has said, ‘Taban is the best mobilizer, and can be dangerous.’ Two people close to the President had warned him before his dismissal as Governor, recommending that Kiir reconcile with him for this reason.119
Taban Deng and Riek Machar are related by marriage, and were on the same side during the 1991 SPLA split and for years later, though at times they ended up in different factions. In 2013, having both been ousted from government positions, they reconciled and were natural partners.120 But we cannot know the extent to which Riek was aware of, and agreed with, everything his comrade was up to.
One source claims that Taban mobilized some Nuer commanders, and particularly officers who were angry about having been taken off the reserve list and were with the President.121 I have not been able to verify this. Sources also claim that Taban and Riek ‘did the numbers’ with regard to the overall percentage of Nuer among SPLA commanders and potential backing in the event of a full confrontation.122 As one senior leader in the Movement from the mid eighties onwards explained to me:
There is no political powerbase in the SPLM without military power. Whoever takes the SPLM, takes the SPLA. The distinction is not there. This implies that you cannot rally political support only. In the SPLM you will automatically also rally military support. The commanders discuss politics. Any political division within the SPLM will eventually lead to a military division. There is no culture to resolve conflicts politically. Hence, they will in the end be resolved militarily.123
This statement is closer to the truth than some may like to believe. It was certainly how Taban and Riek’s activities were interpreted on the government’s side. Meetings with other members of the SPLM Political Bureau were seen in the same light. One prominent minister used the military term ‘regrouping’ to describe the frequent meetings at which planning for press conference and rallies took place, which was seen by him and those close to the President as discussing ‘how they could remove Salva’.124 Several key players on the security side told me that they had evidence of coup plans, but when I asked to see it none was forthcoming. Nor was any such documentation produced during the ensuing court case, which probably means that such plans were never discovered.
Rapid mobilization of Nuer commanders on 16–18 December has been taken as evidence of pre-arranged plans. Peter Gadet has been accused of advance mobilization of the ‘White Army’.125 While prior contacts cannot be ruled out, they appear unlikely, as we shall see. Nuer commanders may, however, have interpreted political rallying as a prelude to military action. A message to several commanders from Taban along the lines that Salva Kiir was trying to humiliate Riek Machar, and that the Nuer should not accept this, could have mobilized some of them. According to some sources, such messages were in fact conveyed.126
Furthermore, at about the time of the 6 December press conference, Taban Deng contacted Nuer members of the Presidential Guards and told them that he, Riek, and others might be arrested, and that they should do what they could to prevent this.127 Sources say that during the week of 10–15 December, Taban spun a rumour that the force at Luri was being deployed to disarm Nuer Presidential Guards.128 Apparently, he was also talking to former bodyguards of Paulino Matip. Several sources from different camps, including two very senior SPLA officers, independently state that rumours of Riek’s arrest from 6 December and the ‘impending disarmament of the Nuer’ escalated the situation. They see this as a major trigger for what happened.
On 15 December at around 3 p.m. a junior officer in the Presidential Guards telephoned Taban Deng with the news that some 120 armed Dinka soldiers (one company) had just arrived at their barracks and made inquiries about the armoury.129 The Guards on duty thought they were about to be disarmed and arrested. According to two eyewitnesses Taban Deng, on the telephone, urged calm and asked a third, a senior general, to do likewise.130 Taban was in Oyay’s house at the time, with Majak and a few others. They requested a top official of one of the security agencies who also came by to alert the President of a serious problem at the barracks and ask for the President’s intervention. He promised to do so. None of the three interviewees are sure whether this was done.
Following the arrival of the armed Dinka soldiers (presumably from Luri), it seems clear that no one at the top level did in fact act to reassure the Presidential Guards that they were not under threat. The exact sequence of events during these early evening hours is uncertain. Later, however, there were a number of phone calls between Taban and Nuer Guards. The security apparatus has referred to transcripts of these conversations as evidence of plans, and they used some in the later court case. These include Taban calling about preparedness, access to arms, and late Sunday evening also directing action. However, from my copies these are selected quotes, in summary form, and hence less accurate.131 As the AU Commission of Inquiry also says, ‘the evidence does not point to a coup’.132
The same SPLA sources allege that the reckless warnings by Taban of their likely arrest triggered the whole situation among the Presidential Guards.133 One account states that a Nuer soldier tried to grab a gun from a sentry, and the sentry himself was killed by another. This action might have been precipitated by an attempt to access the Armoury or by a desire to defend it. A second officer was also shot.134
Whatever the case may be, the atmosphere was like a powder keg, and things got out of control. Dinka and Nuer soldiers started shooting at each other.135 Late evening, following reports from Nuer Presidential Guards, Riek Machar gave the order to break into the armoury.136 Both Dinka and Nuer soldiers took guns from the stores.137 Fierce fighting ensued between soldiers loyal to Salva Kiir and those loyal to Machar. Military intelligence officers were sent in to try to stop the fighting. The source says, ‘it was just chaos; no one knew what the other was doing’.138 The fighting then spread to Bilpam. The Nuer forces took more weapons and uniforms than their own, as they had done elsewhere, reconfirming – in the minds of the President’s people – that a grand scheme was under way.
While the most likely explanation for the outbreak of violence on 15 December was a series of incidents that spiralled out of control, two individuals, one on each side, Paul Malong and Taban Deng, appear to have played important roles in prompting them. They also appeared not to have been entirely unprepared.
SPLA collapses like a house of cards
While the first shots resulted from an escalation seemingly caused by risky decisions and reckless behaviour of a few people, the fighting then exploded. Why was it not controlled?
No message of calm was sent out from either of the two leaders, and the President’s press conference at 11 a.m. did not help. With their subsequent actions, government forces fuelled the fire. As they ventured into the neighbourhoods on 16 and 17 December, where the targeting and mass killings of Nuer were perpetrated, and as forces ‘cleared’ areas presumed to contain Nuer males and civilians, the message was clear to everyone. And as we have seen, during the first few days that followed, key Nuer commanders defected, together with tens of thousands of Nuer soldiers, many of them former militia. The SPLA divided along ethnic lines.
With Juba in turmoil, Bor falling to the opposition, and Bentiu and Malakal following suit, the feeling was of an army imploding, collapsing like a house of cards. The fear that even the capital might be attacked was real. During these critical hours ethnicity trumped everything. Whatever loyalty Nuer soldiers may have had to the SPLA as an institution, or to its commander-in-chief, quickly dissipated. And the forces expected to be loyal to him were fewer in the Greater Upper Nile region than anywhere else.
The most important mobilizing factor was the scale of the killings in Juba. I discussed this later with several of Riek Machar’s commanders; some had been at SPLA headquarters at the time, others in the field.139 They all said the same: they had had no contact with Riek or Taban Deng prior to the crisis. If there had been such contact from Machar’s side, they said, they would not have considered joining him. It was the killings of their own that had mobilized them. ‘When the Nuers were killed like fish, children, women, elderly, and officials and the military were targeted and killed; that is when we realized – this is a tribal war.’ One commander said that it was desperate phone calls from relatives and others in Juba that made him and his force defect and march towards Malakal to fight. He had no contact with Machar for weeks.140 ‘We mobilized ourselves,’ several said.
Sources with intimate knowledge of the SPLA agreed: Riek’s authority over Nuer commanders was limited.141 ‘What Salva Kiir has done, not even the Arabs would do against us; we had to fight back,’ one said. A Nuer politician put it differently, however: ‘This is the war of Paul Malong; Salva Kiir was just dragged into it.’142
Ethnic loyalties were a strong motivator, but personal feelings of marginalization and betrayal probably played a part, fuelled by the desire for revenge. It is also likely that many ex-militia simply followed their old commanders. Former militia leaders defected during the height of the crisis in order to fight the Dinka, the President and the SPLA. This contributed to the surprisingly speedy escalation of the fighting and the SPLA’s losses in the first ten days of the conflict. Reality far surpassed any scenario anyone might have entertained.
The SPLA units most vulnerable to defections were in Upper Nile region and dominated or commanded either by former militia or by Nuer. UNMISS estimated the number of soldiers who defected at 40,000. Since the majority were in the Upper Nile region, the effect on the SPLA was dramatic. The 1st Division in Upper Nile State, the 4th Division in Unity, the 7th Division in Upper Nile (and in particular the units on the west bank of the Nile) and the 8th Division in Jonglei were the most vulnerable, losing much or even all of their strength and a great deal of equipment. Some commanders might have been killed if they had not changed sides: General James Kuong, the commander of the 4th Division in Unity State, is an example; other anecdotal evidence was difficult to verify.143
Forces loyal to the President now had to be moved in from other parts of the country. Recruiting from the regions least affected by conflict resulted in a more Dinka dominated SPLA than in earlier years. The Big Tent had collapsed. It had failed also as a strategy to consolidate peace. The way it had been implemented led to the opposite. Absorption onto the SPLA payroll did not mean integration. With virtually no attempt at addressing grievances politically or personally, loyalties remained with old commanders. This contributed to the domino effect that led to the implosion of the Army.
This development was not of Riek Machar’s making. A very senior SPLA official at the time says, ‘It is Salva who has mobilized for Riek.’144 Once events were in motion, however, Machar willingly and enthusiastically took command of an opposition movement defined more by its grievances than by its membership. He soon took the decision to capitalize on an upsurge of anger and revenge, using the opportunity of massive mobilization. For angry Nuer commanders, their own retaliation and Machar’s decision to topple the President were motivation enough to spur further action. The AU Commission of Inquiry nevertheless concludes that there was an element of coordination ‘that hardly seems possible without forethought’ also on the opposition side.145
The director of SPLA military intelligence, General Mac Paul Kuol Awar helped clarify this important issue. When he gave evidence in the trial of four SPLM detainees, he denied that a coup plot had existed and that this had been the reason for the outbreak of fighting in December 2013.146 Mac Paul was the only major witness not from Kiir’s Bahr el Ghazal inner circle, and his testimony undermined what was widely perceived as an already weak case. Interior Minister Aleu Ateny Aleu, expected to be another key witness for the prosecution, chose not to testify.147
The security dilemma
The Government and the newly minted SPLM/A-In Opposition (SPLM/A-IO), thus blame each other, whether for a coup or for ethnic cleansing or worse, claiming that their reactions only were defensive responses to aggression of the other. In my thorough research on these hypotheses, I have not found enough evidence to indict either side for sole blame.
Many commentators have rejected both sides’ stories and focused instead on two other explanations for the eruption and escalation of civil war. On the one hand, there has been a tendency to resort to simplistic ethnic explanations, reminiscent of former colonial powers’ description of events in tribal terms: this is a war between the Dinka and the Nuer. On the other hand, the focus has been on the two intransigent leaders, whose selfish ambitions and greed for power have driven the country to total disaster. Neither explanation holds water.
First, the crisis was primarily a political one, and not driven by ethnic agendas. Ethnicity became a factor exacerbating the crisis, first through events in Juba, and later through revenge in other places in the Upper Nile region. Ethnicity and its representative structures (‘tribes’) emerge as the consequence rather than the cause of conflict.148 And while greed in terms of power and resources most likely played a part in escalating the political crisis, it does not explain the explosion of violence.149 While these factors probably contributed to the obstinacy of both leaders and prolonged the crisis, they do not answer the question of why the mutual escalation was not stopped.
In an environment without a culture of resolving conflicts politically, no mechanisms within the party to do so, and a polarized environment and virtually no credible and impartial mediators, the stakes were high. Instead of choosing the route of reconciliation and compromise, each side counted on the other to succumb to political pressure through a showdown. Preventive action was taken, particularly by Paul Malong and Taban Deng, in the expectation that the other side would resort to violence. What originally could be characterized as a political crisis very quickly became a security crisis.150
Another explanation for these developments may be found in what has been called ‘The Security Dilemma’ in an article by Barry R. Posen:151
what one does to enhance one’s own security causes reactions that, in the end can make one less secure […] All fear betrayal […] they are unaware that their own actions can seem threatening […] [nevertheless] […] The nature of their situation compels them to take the steps they do.
These actions can trigger a preemptive response. Posen applies this idea to situations where central authority collapses in new states that are governed by a multi-ethnic leadership. Indeed, those who emerge as leaders of any group and confront the task of self-defence for the first time will doubt that the group identity of others is benign. The result is that groups see each other only as threats, and will ‘arm’ – militarily and ideologically – against each other. These dynamics can produce incentives for preventive war, and can even lead to a drive for security so great that it produces near-genocidal behaviour.152
Salva Kiir and his closest allies were convinced that Riek Machar and his network were preparing to act militarily. There are no traditions in South Sudan for violent military coups; more likely would there have been a mutiny, with those loyal to the President losing control. This probably explains the increase in Presidential Guards through the Luri forces, loyal to the President largely on ethnic grounds, and the military recruitment. With the press conference on 6 December, the security forces in Juba were on maximum alert, and Presidential Guards and other units were, according to some sources, armed at critical points all over town. There had been unconfirmed reports about a potential showdown. They remained at this level in the run-up to the meeting of the National Liberation Council.
While we international observers noted increased military presence, the ethnic identity of the soldiers was difficult for us to detect. Sources have indicated that the armed soldiers deployed in town were primarily Dinka.153 At this particular time, tensions between the President and Riek Machar were at their peak. As one could expect, Dinka domination among the forces in the streets created concern among Nuer commanders and soldiers in the barracks.
The President and his allies were so suspicious of Riek Machar and his capacity to mobilize the Nuer that they took precautionary measures. As a credible Sudanese source with intimate knowledge of the dynamics at the highest levels in Khartoum said to me a few months into the crisis: ‘The people around Salva Kiir have a security mentality; Juba has become just like Khartoum.’154 The behaviour of Taban Deng probably gave them reason to believe that something was afoot. Antagonism between the two factions, and historical grievances between the two communities were so great, that every move was interpreted in the worst possible way. ‘They must assume the worst, because the worst is possible.’155
As a credible source has said, in the SPLM you cannot rally political support only; you will be assumed also to rally military support. Any political rift within the SPLM will imply a military rift.156 Engaging in risky behaviour, as all the factions of the SPLM did at this point, was very dangerous at a time when Juba was like a tinder box. Already, Juba was rife with rumours that the Dinka and Nuer were going to fight.
It is likely that both sides made their own calculations of support in the Army, should things deteriorate. This might explain the additional Dinka forces deployed to Juba, and particularly to the Presidential Guards. On Machar’s side, calculating the numbers of commanders presumed to be on ‘his side’ could for example be used as political pressure against the President and chairman of the SPLM at the right time, and especially when well planned. What is clear, however, is that such calculations in themselves can contribute to escalation. One can lose control over what follows. In a tinder-box environment, any incident can light the fire. And in Juba on 15 December an incident in the Presidential Guards did.
Exactly what happened during the night of 15–16 December can only be clarified through independent investigations. The same is true of what lay behind the targeted ethnic killings, massacres and atrocities by Dinka security forces, in particular 16–18 December.
Game changers in battle
More thorough analysis will also be needed to achieve full understanding of how and why the violence spread beyond Juba. One simple but important factor must be mentioned: the seasons. If these developments had happened at another time of the year, it is unlikely that we would have seen the scale, speed, and duration the conflict reached. That the crisis erupted when the dry season had just started made it possible for fighting to spread quickly.
Evoking unavenged killings is one way of mobilizing a clan or ethnic group in Nilotic cultures. After the massacres in Bor by Riek Machar in 1991, which killed at least 2,000 people, no one had been held accountable. Nor had there been revenge. ‘1991’ was a terrible nightmare recurring in the minds of many SPLA cadres. It was perceived as a mobilizing factor for the violence in Juba during these December days, and a reference point for a lot of public rhetoric, from the President’s speech to the National Liberation Council on 14 December, in the press conference on 16 December, and in other commentary from leading government figures during the crisis. It was repeatedly said that Riek was at it again. This accusation may itself have incited ethnic violence. Whether this was intended or not is difficult to tell. That ‘1991’ seemed to have become code for targeted killings of Nuer among some Dinka forces seems nevertheless to be the case.
I discussed ethnically motivated killings with President Kiir many times, expressing grave concern. He agreed, underlining that he had not used ethnicity as a mobilization tool against the opposition. Kiir stated that if he had done that, the whole of the Dinka of the Bahr el Ghazal would be on the move, and there would be total havoc in the country. This was a conflict fought by government forces against a rebellion, and nothing else, he claimed. But the Dinka recruits from Luri who were devoted to his protection, and the rhetorical use of ‘1991’ were still important mobilization factors on ethnic grounds. Revenge was an important reason for the massive response on the Nuer side, as we have seen.
Together, the personal grievances and the gravity of atrocities seem to have constituted a groundswell of movement from below. Riek Machar certainly took opportunistic advantage of this. But he was not himself the game changer. To quote one prominent Nuer:
What happened in Juba, the killings, changed everything with all the Nuer. Riek was not a popular figure, and would normally not command support from all the factions of the Nuer. But with the killings in Juba, and the fact that Riek himself almost got killed by the Government forces, the sympathies went in his direction, and the Nuer factions joined in and rallied behind him.157
The scale, speed and duration of resistance cannot be explained by Machar’s actions. There is no way he would have the personal gravity and unifying force among the Nuer to make this happen at his own merit, or the capacity to orchestrate such enormous opposition, leading to massive defections in Bor, Unity and Malakal in just a few days.158 As he was fleeing for his life from the SPLA, whose units were literally hunting him down, he was not in a position to direct multiple actions in far-flung locations – even if he had wanted to.159 According to several credible sources, Machar only narrowly escaped government forces outside Juba and was almost ambushed by the SPLA elsewhere within the first 48 hours.160
The Bul Nuer commander of the 8th Division in Jonglei State, Peter Gadet, was assumed to play a more important role.161 When Bor fell so quickly to opposition forces it was presumed that he had orchestrated it in advance, in cooperation with Machar. Gadet’s absence from the commanders’ conference in November had even then been seen as suspicious, and he had to be persuaded to come to Juba.162 Government sources also refer to reports from the commander in Waat indicating that Gadet engaged in suspicious activities there; it was later assumed that he had started mobilization of the White Army at this time.163
Others familiar with Gadet’s thinking say that his reluctance to come to Juba resulted from his escape from an ambush the previous April. He feared another attempt on his life.164 But there was no love lost between Gadet and Riek Machar,165 and Gadet and Taban Deng were also mutually hostile.166 It must be noted, however, that on 20 November 2013 Taban and Gadet are alleged to have had a private meeting in the margins of the commanders’ conference. I have not been able to confirm what they discussed, but afterwards Taban started to talk again about arrests, and of the likelihood that Gadet would be arrested.167
Nuer members of the Auxiliary Police were the first opposition forces to enter Bor town, and their actions were reportedly decisive in determining what happened later. Their move appeared to be spontaneous, as UNMISS reports also indicated. According to eyewitness accounts, Gadet went to Bor to try to calm things down.168 It was only later, after the events in Juba on 16 December, that he decided to switch sides, team up with the police, and take over the town.169 Whether this was part of a plan or a spontaneous move, as several of Gadet’s colleagues – as well as foreigners in touch with him at the time – claim, is now not possible to verify.170 According to the account of people in Riek Machar’s convoy, and Machar himself, driving to Bor – after a very hazardous journey, much of it by foot, – developments in the state capital were as surprising to the former Vice President as to the government.171 Whatever the case may be, upon arrival on 24 December Machar met Peter Gadet, Hussein Maar (the Nuer former deputy governor of Jonglei, who had sought refuge in the UNMISS base), and a number of other government officials. Hussein and the former County Commissioner Goi Yol joined Riek Machar’s convoy heading north from Bor. In Mathiang, the home village of Kuol Manyang,172 they were caught in an ambush on 25 December, and only narrowly escaped before proceeding to Gadiang.173
While Gadet’s actions were important, it appears that the government’s forces’ own actions precipitated reactions that united the Nuer behind Riek Machar, and made it possible for him to assume command. But as one of Machar’s old compatriots from the Nasir faction (who did not join him now) says: ‘The mistake Riek made was to take over a rebellion he didn’t plan. And when they committed atrocities, he was responsible.’174
There were two branches of the White Army, one of the Lou Nuer in the north-western counties of Jonglei State, and the other in the south-eastern parts of Upper Nile. Although the White Army originally were Nuer armed youths mobilized by Riek Machar in the early 1990s, it was not owing to his efforts that they gathered now. According to credible sources, the attacks by Murle armed youths in the county of Twic East, as well as in Lou Nuer areas in past months, had led to a major mobilization of Lou Nuer youths and the White Army at the beginning of December.175 One source, a young Lou leader, says that many armed youths were already gathered at Gadiang when Riek Machar arrived. They were several thousand, and more were on the way.176
From Riek Machar as well as from other sources, it has been made clear that extensive discussions took place with the White Army leadership over several days.177 On command and control, discussions were difficult, but Gadet got them to agree to his military tactics despite their reluctance.178 The White Army leaders in Jonglei also agreed to refrain from attacking women and children and innocent civilians. That these arrangements helped initially seems clear,179 and more so than with the armed youths in Upper Nile State.180
I doubted that Machar would be able to control the White Army and warned him about the prospect of their attacking civilians, and women and children in particular.181 The opposition leader seemed unrealistic about his capacity to keep them in check, and claimed that they would behave as they moved southwards towards Bor, which at this point was part of the plan. The mobilization of thousands of angry armed youths and the White Army was a factor to be reckoned with, and soon the stories began that their next target would be Juba. For them, loot was a major motivator beyond the desire for revenge, and it did not take long before codes of conduct were violated.
Unexpected allies
Foreigners made contingency plans for worst-case scenarios for Juba, including evacuation of all – or all non-essential – staff. Some had already completed evacuations in late December, others followed suit. For UNMISS it was critical to remain, albeit with adjusted numbers, but the Mission revised its own plans and prepared for the worst.
The speed and scale of the violence and the mobilization of the opposition led the government to request support from the outside.182 The prompt deployment of the UPDF on 19 December, to Juba and beyond, seemed to be based on the SPLA’s fear of imminent defeat. While Ugandan protection of strategic installations in Juba received IGAD’s blessing, the engagement of the Ugandans in Bor and surrounding areas was more controversial.
Without the UPDF deployment, including its air capabilities, it is not clear what would have happened. It was a game changer. The most decisive factor from a military perspective was the Ugandans’ use of heavy artillery, which effectively dissuaded Peter Gadet and the White Army from advancing on Juba.183 Ugandan support therefore gave the government time to stabilize the situation and regain the initiative. Following the attacks on Bor and operations in some other places, but with no supplies and nothing more to loot, the White Army returned to their respective communities. White Army elements operating in Upper Nile State similarly withdrew, at least for now.
Two other factors, not as decisive, were still important. Two militia leaders who had recently been reintegrated into the SPLA remained loyal to President Kiir. This was quite surprising, given the history of their relations. Both the SSLA militia leader Bapiny Monytuiel and the commander of the Shilluk militia Johnson Olonyi signed an agreement with the government only six or eight months prior to the crisis. The former SSLA’s control over important areas in Unity was essential for the government’s ability to hold the state. The new Governor, Joseph Nguen Monytuiel, had broken ranks with Riek Machar, with whom he had previously been close, and joined the government in April 2013. It is likely that he did his utmost to ensure that his brother, the newly integrated SSLA militia leader Bapiny, also remained on side. Defection by the former SSLA could have had a huge impact in Unity State.
Complicating the alphabet soup, the South Sudan Democratic Movement and Army (SSDM/A), Upper Nile faction, often called Johnson Olonyi’s forces or the Shilluk militia, accepted the President’s amnesty after significant pressure, not least from the Shilluk reth (king), in June 2013. But while Olonyi returned to Juba, his troops stayed in the field as negotiations of the terms dragged on.184 The group remained loyal to the government for more than 16 months into the conflict, until an incident perpetrated by the SPLA caused them to jump ship. What then ensued, the loss of Malakal to the opposition, showed what might have happened if Olonyi’s forces had joined the opposition earlier.
If the White Army had moved on Juba, and the militia groups had gone over to the opposition, the SPLA could not have held. This would have been a dramatic turn of events. That this did not happen, and the UPDF remained on South Sudanese territory, no doubt prevented havoc in Juba. But it also permitted the conflict to continue unabated. It was prolonged not only with the availability of arms on both sides, but mobilization through the ethnic card, with stereotyping, scapegoating, dehumanizing rhetoric and creating a sense of victimhood through ‘reproduction’ of grievances.185
In the battlefield, the two sides continued to pursue a tit-for-tat approach, often avoiding direct confrontation, focusing on control of strategic locations in the Upper Nile region. Civilians bore the brunt. As so many times in South Sudanese history, the security forces, whatever their allegiance, did not act as protectors of their people, but as predators.
That innocent South Sudanese should go through these nightmares once again, after they had finally won their independence, was bad enough. But never before, even under Khartoum’s leadership, had anyone seen such atrocities committed by South Sudanese against South Sudanese. Never before had the towns of Bor, Malakal and Bentiu seen such destruction. Thousands and thousands had been killed. Never before had the sanctity of churches, mosques and hospitals been violated in this way, not even under attacks by Khartoum-recruited militia.
On Independence Day two years before, the President had said:
We the people of South Sudan have experiences what it is to be a refugee. We hope that this has been our last war and that our people will never again have to cross our borders in search of security.186
By July 2014, 1.5 million were displaced. Hundreds of thousands were refugees in Kakuma and across the Ethiopian border and, even more shockingly, in Sudan.
The country was also at risk of breaking another record, of seeing the worst famine in its history unfold. Predictions were clear: in 2014 close to 4 million people were estimated to be in urgent need of food aid187 because of this manmade disaster.
The question was: would any of the antagonists be ready to come to an understanding? What were the incentives for agreement? Would they – in the end – be willing to put their people and the future of their country, and peace, above their own personal interests?