There is little debate among scholars and policymakers that intrastate conflict is currently the most prevalent type of armed conflict. In fact, of the thirty-one armed conflicts that were ongoing in 2012, all but three occurred mostly within the boundaries of one state. Conflict recurrence is another dominant feature of modern armed conflicts, with half of these intrastate conflicts seeing a return to violence after an earlier agreement was signed between warring parties. On the African continent, the numbers are similar. In 2012, of the twelve ongoing conflicts in Africa, all but two were intrastate conflicts, and half of these had seen at least one relapse into violence after an agreement was signed (Kreutz Reference Kreutz2010; Högbladh 2011; Themnér & Wallensteen 2013). These statistics tell the story of stalled and failed peace processes, but they do not point to potential causes for these failures.
Why have these peace processes not achieved lasting peace? The problem does not seem to be getting combatants to the negotiating table; rather, it is making sure that all of the key combatants are at the table, and keeping them committed to staying there. This need to recognize the key combatants in an armed conflict is usually overlooked by researchers, who often treat rebel groups as unitary actors. The empirical evidence, however, shows that of the five African conflicts with at least one failed agreement that were active in 2012, all of them had multiple rebel groups fighting against the government and each other (UCDP 2012). Each of these rebel groups can decide to participate in or withdraw from a peace process at any given time, or never enter the process at all. Understanding why a rebel group decides to spoil the peace process is essential if policymakers hope to successfully end a violent conflict.
I argue that changes in a rebel group’s grievances, and its assessment of the costs and benefits of violence, lead to changes in a leader’s decision to participate in or spoil a peace process. Specifically, I argue that an increase in a group’s grievances, a decrease in the costs associated with violence, and an increase in the benefits of violence will lead to a decision to spoil a peace process. In order to explore this argument, I contrast the Sudan People’s Liberation Army’s (SPLA) decision to spoil the 1997 Comprehensive Peace Agreement (CPA) with its decision to sign and abide by the 2005 CPA.Footnote 1
Exploring Theories of Conflict Recurrence and Peace Processes
In order to explain rebel group leaders’ decisions during a peace process, I turn to studies of actors’ decision-making during conflicts, most of which focus on three factors that influence these decisions: uncertainty, grievances, and the costs of violence. Researchers who focus on uncertainty as the major cause of conflict recurrence argue that actors’ uncertainty about the future, and fears of retaliation, make them unwilling to trust their opponents, and this lack of trust explains why actors continue fighting. To overcome this uncertainty, researchers offer several policy prescriptions, the most popular being the inclusion of third-party guarantors to monitor and enforce peace agreements (Walter 2002, Reference Walter2009).
Researchers who examine actors’ grievances argue that conflict occurs because of a gap between what individuals believe they are entitled to and what they believe they can achieve (Gurr Reference Gurr1970, Reference Gurr1993; Humphreys & ag Mohamed 2005; Sambanis Reference Sambanis2001). This gap is what leads to group grievances, which Gurr and Moore define as “widely shared dissatisfaction among group members about their cultural, political, and/or economic standing vis-à-vis dominant groups” (1997:1081). Regarding the cost of violence, Zartman (Reference Zahar1989, Reference Zartman, Darby and MacGinty2003) explains that actors will decide to end a conflict only when there is a “mutually-hurting stalemate”—that is, a balance of military power between two relatively equally matched opponents that makes the use of violence too costly. In a related theory, Monica Toft (2010) argues that peace processes have a successful outcome only when actors are threatened with “mutual harm” if they resume violence. She suggests reforming the security sector of the state (e.g., police, military, intelligence, court system) so that it is better equipped to punish perpetrators of violence.
These motivating factors contribute to an understanding of rebel group leaders’ decisions to spoil a peace process, but when taken individually, each focuses too much on one factor at the expense of others. For instance, empirical studies show that the presence of a third-party guarantor increases the likelihood of conflict termination. These studies, however, are unable to explain why conflicts in which third parties have been engaged to facilitate and guarantee peace negotiations can still drag on for a period of months and years; for example, the armed conflict (and later genocide) in Rwanda continued despite the Arusha Accords’ backing by the United Nations and others. Explanations based on the existence of political, social, economic, or other grievances cannot account for groups that have grievances but do not take up arms; one example is the Movement for the Survival of the Ogoni People in the Niger Delta region of Nigeria. Finally, explanations that focus only on the costs associated with violence cannot explain decisions on the part of rebel group leaders to begin or continue perpetrating acts of violence even against a much stronger opponent, as in the case of le Mouvement des Forces Démocratiques de la Casamance in Sénégal or the secession of the Republic of Biafra from Nigeria.
When taken as part of a more comprehensive explanation of rebel group leaders’ decisions during a peace process, each of the three motivational factors listed above has something important to say about a group’s decisions during a peace process. Specifically, combining different motivational factors can shed light on rebel group leaders’ decisions to spoil or participate in a peace process. The idea of examining the role of rebel groups as spoilers was proposed by Stedman (Reference Stedman1997), who focuses not only on groups’ motivations for perpetrating violence, but also on the particular types of actors that use violence during a peace process. These actors, known as spoilers, are defined as “leaders and parties who believe that peace emerging from negotiations threatens their power, worldview, and interests, and use violence to undermine attempts to achieve it” (1997:5).
The focus on spoilers was a clear and important shift in the study of conflict, but it is not without its critics. These criticisms do not argue that examining spoilers is not important, but that assumptions about the types and motivations of spoilers should be more nuanced. Darby (2001), Zahar (Reference Zahar, Darby and MacGinty2003, Reference Zahar2010), and Greenhill and Major (2006) argue that Stedman’s typology of different spoilers places them in a particular category, but it does not explain how or why groups move between the categories. Specifically, why do groups labeled as “total spoilers”—the groups least likely to sign agreements in good faith because their goals are absolute and unchanging—eventually become willing to negotiate at a later point in a peace process? Stedman argues that groups are often mislabeled as “total spoilers” for political reasons, and that a group can be labeled as such only once an agreement has been signed; “in the absence of an agreement,” he argues, “the concept of spoiler should not apply” (2003:108). Zahar (Reference Zahar2010) argues that the problem is not that groups are mislabeled, but “that there are no fixed spoiler types” (2010:266). I agree with this argument and add that based on Stedman’s definition of “spoiler,” specifically his reference to negotiations, there is not a clear reason that the label “spoiler” cannot be affixed to groups that use violence to disrupt a peace process before an agreement is signed (and even if no agreement is signed). If negotiations are considered only one phase of a larger peace process, which may or may not include a signed agreement, then a group is capable of using violence to undermine the process at any point after negotiations begin. For this reason, I use the label of “spoiler” for any group that uses violence during any phase of a peace process.Footnote 2
Zahar (Reference Zahar, Darby and MacGinty2003, Reference Zahar2010) also critiques the spoiler typology for its failure to consider the structural context within which spoilers operate. As such, she examines leaders’ cost–benefit analysis and how this analysis (and the resulting decisions) changes over time. She identifies rebel group capability and opportunity to use violence as the key factors that influence group decisions about whether to spoil or participate in the peace process. Because capability and opportunity both change relatively often, Zahar is able to account for the fact that actors’ decisions to spoil change over time, something many other studies are unable to do. The rest of this article lays out a comprehensive theory of rebel group leaders’ decisions in order to explain why these decisions may change during an ongoing peace process.
Rethinking Rebel Groups and Conflict Termination
Like Gurr, I argue that a comprehensive theory of actors’ decisions during conflict must examine their grievances, and like Zahar, I argue that such a theory must also examine the role of structural factors. I argue, specifically, that changes in rebel group grievances, and the costs and benefits of committing violent acts, lead to changes in leaders’ decisions to spoil peace negotiations.
As discussed above, group grievances are specific complaints that a rebel group has against the government. To identify specific rebel group grievances, I have relied (like Gurr) on the “Minorities at Risk” (MAR) dataset compiled by the University of Maryland’s Center for International Development and Conflict Management (2009).Footnote 3 MAR identifies several types of grievances, including political, economic, and social/cultural grievances. For each type of grievance, I examined whether rebel groups have made public statements on the particular topic or whether other evidence exists showing discontent in the general public. For example, I identified whether the group’s leaders have called for greater political rights, or whether citizens have demonstrated against the lack of the freedom of religious belief and practice. I expected that as group grievances increase, leaders are more likely to undermine the peace process because of frustration with the government’s unwillingness or inability to address these grievances.Footnote 4
Regarding cost–benefit analysis, there are several factors that influence this analysis of whether to spoil or participate in a peace process. Building on the theories of Zahar, Zartman, and Toft, I identified three factors that affect the cost of spoiling a peace process: (1) the rebel group’s capacity to use violence; (2) the potential harm from retaliatory violence; and (3) possible exclusion from an existing peace process. Rebel group capacity determines whether the group is able to carry out violence. Using Zahar’s operationalization of rebel group capability, I measured changes in group capacity by identifying reported changes in money, weapons, and third-party financial support. To this list I added physical resources, including the number of group members and whether neighboring countries allow groups to use their land for bases. I expected that an increase in rebel group capacity would lower the costs of spoiling the peace process and therefore would make a leader more likely to do so.Footnote 5
Along with changes in rebel group capacity, I analyzed the second cost of committing violent acts: the threat of harm from retaliatory violence. In most violent conflicts, rebel groups know that any action on their part will elicit a response from the government as well as from any external actors involved in the peace process. In order to measure the threat of harm from the government, I observed past government repression, specifically the following indicators from MAR: arrests, imprisonment, kidnapping, torture, disappearances, and murder. I also observed new government military campaigns and increases in other combat-related actions as indicators of the threat of harm from the government. To measure the potential harm from external actors’ countermeasures, I used Zahar’s “commitment of foreign actors” indicator. Zahar explains that any credible commitment by external actors involved in a peace process must include a publically stated willingness to “oversee the implementation of the peace [process] and decisively deal with would-be spoilers” (2010:271). I expected that as the threat of retaliation decreases, rebel leaders would be more likely to be spoilers.
The third category of factors affecting rebel groups’ costs of spoiling includes all of the adverse effects that come from being excluded from an ongoing peace process. This category of costs does not affect all rebel groups equally. According to Stedman, Zahar, and Khatib (2006), the threat of losing that seat (for groups that already have a seat at the negotiating table) is greater than the threat of not getting a seat (for groups that have already been excluded from the process). For this reason, groups that have been included in the peace process face a higher cost of spoiling. Exclusion from existing peace talks and the possible agreements that might result from them leads, in turn, to exclusion from any power-sharing governments as well as a loss of material benefits (e.g., aid from foreign donors, revenue sharing from natural resources) that might be gained from peace. These consequences of being excluded from peace negotiations obviously do not exist for rebel groups that have been outside of the peace process to begin with. When calculating the costs of spoiling a peace process, outside rebel groups must account only for their group capacity and threat of harm. For this reason, I expected that the costs of being excluded from an ongoing peace process would make leaders of included rebel groups less likely to be spoilers.
The potential benefits of spoiling a peace process for outsiders include a possible military victory or else eventual inclusion in the process. According to Zartman’s concept of “ripeness,” all parties to an armed conflict consider the potential benefit of a military victory, which would lead to an optimal outcome for the victors. In order to measure the potential for military victory, I looked at a group’s battlefield victories and territory won. These indicators, which come from Greenhill and Major (2006), look at the outcomes of a group’s immediate past military actions to identify its perceived potential for future military victory.Footnote 6 Unlike a potential military victory, however, the benefit of using violence to be included in an existing peace process applies only to those rebels who are outside of the current process. As such, I expected that the benefit of military victory would make both included and excluded groups more likely to act as spoilers, and that the benefits of being included in the current peace process would make outside groups more likely to act as spoilers than inside groups.
Hypotheses and Methods
The above argument and expectations generated five hypotheses. The first hypothesis deals with the impact of changes in rebel group grievances on leaders’ decisions to spoil a peace process.
H1: If a rebel group’s grievances increase, a leader is more likely to spoil the peace process.
The next three hypotheses address the factors that influence the cost–benefit analysis of spoiling that applies to both included and excluded rebel groups.
H2: If a rebel group’s capacity to use violence increases, a leader is more likely to spoil the peace process.
H3: If the threat of harm from retaliatory violence against the rebel group decreases, a leader is more likely to spoil the peace process.
H4: If a rebel group’s potential for military victory increases, a leader is more likely to spoil the peace process.
The final hypothesis addresses the different cost–benefit calculations for groups that are included in the peace process versus groups that are excluded.
H5: If a rebel group is not part of peace negotiations (external), the leader is more likely to spoil the peace process than the leader of a rebel group that is part of peace negotiations (internal).
In order to explore these hypotheses, I examined two key events in the twenty-two-year civil war between the Government of Sudan (GoS) and southern Sudan—the 1997 Comprehensive Peace Agreement (CPA) and the 2005 CPA—and compared the peace processes that led to these agreements. This “process trace” allowed me to identify the actual process by which changes in the explanatory variables led to changes in rebel group decisions about whether or not to spoil the process. The goal of the process trace was to collect and analyze data that provide evidence to explain why the SPLA spoiled the 1997 CPA but signed and abided by the 2005 CPA.
I chose these two peace agreements for comparative analysis because examining two different peace processes in the same country (and the same conflict) allowed me to hold significant variables from alternative explanations for conflict termination relatively constant, including natural resource endowment (greed explanations based on the goal of material gain) and ethnic fractionalization (heterogeneity). Despite important similarities, the 1997 and 2005 CPAs had very different outcomes, with the former never recognized by the original SPLA and the latter signed by the SPLA and all other southern groups operating in 2005. More important, by studying the different decisions made by the same rebel group leaders, I was able to highlight the external changes that led the same group to make two very different decisions.
The SPLA has been fighting the GoS for either autonomy or independence since its founding in 1983, with the most often cited reasons being the racial, religious, and ethnic differences between the peoples of northern and southern Sudan. The main source of the South’s grievances is the fact that these differences are directly related to the distribution of power and wealth in Sudan. Another significant grievance grew out of an event that would change Sudan forever: Chevron’s discovery of oil in Bentiu, southern Sudan, in 1978.Footnote 7 Though this discovery should have been the beginning of a better economic future for the people of the region, they became increasingly aware that the government had little intention of allowing them to reap the benefits of the discovery. A third key grievance for the people of the South was the GoS’s announcement in September 1983 that the entire country would follow Shari’a law. Though the SPLA had been founded two months earlier, the writing had been on the wall since 1980, when President Numairi’s increasingly Islamist government began merging the existing legal system with Shari’a law. Given that the vast majority of citizens in southern Sudan are Christian or animist, this “Islamization” was an important catalyst for the founding of the SPLA (Post 2005; Lesch 1998).
The 1997 Comprehensive Peace Agreement
From its founding in 1983 the SPLA, with help from Ethiopia (and later Eritrea and Uganda), gained the upper hand in almost all of southern Sudan. By the early 1990s, however, the SPLA began facing internal strife, and 1991 saw the group’s first official split. This event, along with later splits, led to fighting between the SPLA and its various splinter groups, which in turn resulted in the loss of territory that the SPLA had gained in its struggle against the GoS. Nevertheless, the SPLA saw increasing military success by the mid-1990s.
As the SPLA was gaining ground in early 1996 in southern Sudan, two splinter groups, Riek Machar’s South Sudan Independence Movement (SSIM) and the SPLA-Bahr al Ghazal (SPLA-BAG), were in the midst of peace talks with Khartoum. In April, the two groups signed a peace agreement with the GoS. Though not the first agreement in this war, it was the first in a series of peace agreements that occurred between 1996 and 1997. Even though it was less comprehensive than later agreements, the April 1996 agreement marked the first time key issues in the conflict were addressed, and as such it was the beginning of the peace process that led to the 1997 CPA. For reasons discussed below, the SPLA remained outside of the continuing peace talks among the signatories of the April 1996 agreement, and it continued its military campaign against the GoS throughout the rest of 1996. In April 1997 the GoS signed the Khartoum Peace Agreement, a CPA between the government and several SPLA splinter groups: the SSIM, SPLA-BAG, the Nuba SPLA, the Equatoria Defense Force (EDF), and the SPLA-United (signed retroactively in September 1997). This peace agreement received the designation “comprehensive” because it clarified various clauses from previous agreements, provided a skeleton plan of implementation that was lacking in previous agreements, and addressed some of the Southerners’ key grievances against the GoS (Government of Sudan 1997).
Hypothesis 1: Rebel Group Grievances
In the years leading up to the 1997 CPA, there were few increases in the SPLA’s grievances against the government. In fact, the SPLA’s grievances appear to have decreased, which should have led them to sign and abide by the 1997 CPA. In 1996 the SSIM and other SPLA splinter groups signed a peace agreement, and this agreement laid the groundwork for the 1997 CPA. The CPA considered issues of self-determination, the right of Southerners to hold political positions in the federal government, and South Sudan’s autonomy within the Sudanese federal system. On the surface, then, it would seem that most of the SPLA’s key grievances were resolved. But this CPA actually contributed to an increase in SPLA grievances because it did not clearly lay out a plan for implementing these changes, leading to doubt over whether the GoS planned to do so. For example, even though the CPA proposed a referendum on self-determination for the South, the time frame for such a referendum was tentative at best, and the logistics for how it would be carried out were not specified. Though the SPLA was clear about its desire to remain part of a united Sudan, it was adamant that the people of southern Sudan should have a say in their future. Based on this CPA, though, the SPLA did not have much faith that such a referendum would actually take place.
Second, the CPA did not address the issue of whether Southerners would be subject to Shari’a law (which was both a political and social-cultural grievance). The GoS’s “Islamization” in the early 1980s had been one of the catalysts that led to the creation of the SPLA in the first place and the beginning of the civil war. The SPLA refused to sign an agreement that did not address this issue; as such, it did not sign the 1997 CPA.
Hypotheses 2 and 3: Rebel Group Cost of Using Violence
Unlike group grievances, the cost to the SPLA of using violence underwent major changes in the period leading up to the 1997 CPA. Specifically, the SPLA’s capacity to commit violent acts increased because of help from other countries and an alliance with opposition groups from the North. By the mid-1990s Sudan’s relations with its neighbors had soured, damaged by Uganda’s and Ethiopia’s accusations that Sudan was helping rebels in their countries. Sudan’s ties with Egypt were also worsening because Egypt accused Sudan of involvement in an attempted assassination of President Mubarak, and the international community condemned Sudan for harboring three men accused in the crime. For its part, Sudan accused Ethiopia, Eritrea, and Uganda of providing resources and safe haven to the SPLA. At various points during the mid- to late-1990s, all three countries were implicated in providing weapons, equipment, money, land, and possibly troops to the SPLA. During this same period the United States provided not only $20 million of military equipment to the SPLA’s reported allies, but also a team of covert ground troops to southern Sudan. In addition, the SPLA gained resources from the GoS weapons and military equipment it captured in its military campaigns. All of these resources greatly increased the SPLA’s ability to use violence against the GoS (Human Rights Watch 1999).
At the same time that it was receiving substantial resources from the U.S. and neighboring countries, the SPLA began strengthening its political and military alliances with opposition groups from the North. Since 1990 the SPLA had been loosely tied to an umbrella group, the National Democratic Alliance (NDA), but it sought to reinforce the alliance by working out some of the political differences between the SPLA and the northern opposition groups. Once these differences were worked out, the SPLA and NDA began military operations against the GoS. In 1996 the SPLA’s leader, John Garang, became the leader of the Joint Military Command of the NDA, and “for the first time a southern commander gained operational responsibility over northern troops” (Lesch 1998:202). With the NDA alliance solidified, the SPLA’s capacity to use violence against the GoS was at an all-time high.
Along with the SPLA’s capacity to use violence, there were changes in the threat of retaliation from the GoS. Despite a major SPLA military campaign that began in the mid-1990s, the GoS actually retaliated very little. In fact, there were only eight reported GoS acts of aggression against the SPLA from 1995 to 1997, and only four of these attacks were successful. Though the GoS was able to capture two towns, its losses were important for the SPLA because it showed that the GoS was unable to militarily defeat the SPLA. With the last of these GoS defeats coming less than a month before the 1997 CPA, the threat of retaliation for spoiling the peace process and the agreement was small and decreasing.Footnote 8
Hypothesis 4: Rebel Group Benefit of Using Violence
Regarding the SPLA’s potential for military victory, there were major changes before and during the peace process that led to the 1997 CPA. The SPLA’s military operations as part of the NDA threatened Khartoum more than the SPLA was able to on its own, bringing the war several hundred miles closer to the capital. Beyond Khartoum, the NDA was in control of towns on Sudan’s border with Eritrea and Ethiopia, and it enjoyed support not only from northern opposition leaders, but also from Ethiopia and Eritrea. By early 1997 the SPLA controlled the vast majority of southwestern Sudan, and just one month before the signing of the CPA the SPLA launched a military operation allowing it to regain control of more than twenty towns in the southeast. By the time the CPA was signed, the SPLA and NDA “controlled a 114 square kilometer area with 29 towns and villages along the Red Sea coast, with 92,000 residents, plus 86 towns and villages in Hamesh Koraib [Kassala province in northeastern Sudan] with 360,000 residents” (Lesch 1998:204). Based on these and other victories throughout southern and eastern Sudan, the SPLA had strong reasons to believe that military victory was not only possible, but likely.
Hypothesis 5: Rebel Group Position in the Peace Process
The final explanatory factor in rebel groups’ decisions to spoil the peace process is their position in the process, which can determine whether they will receive future rewards as part of a peace agreement. Groups that are included in a peace process are expected to be less likely to spoil it for fear of losing the benefits of inclusion, while groups excluded from a process are more likely to try to spoil it in an effort to gain the potential benefits of being included in the future. In the year leading up to the 1997 CPA it was clear that the SPLA was not part of the peace process. In fact, in early 1996 the GoS began peace talks with leaders of two SPLA splinter groups, leading to the peace agreement in April 1996. As mentioned above, this agreement failed to address several issues—including religious freedom—and was rejected by the SPLA. For this reason, the SPLA refused to participate in the peace talks that continued between Khartoum and the splinter groups. When the 1997 CPA was signed, then, the SPLA was squarely outside of the peace process.
The 2005 Comprehensive Peace Agreement
Southerners’ hopes for peace began decreasing just three months after the signing of the 1997 CPA when Sudan’s legislature, the National Assembly, ratified a different version of the agreement. The July 1997 changes to the CPA had the net effect of reducing the powers of the Coordinating Council of the Southern States, the main governing institution of the newly autonomous southern Sudan. In addition to these external changes to the agreement, there were internal problems in the Coordinating Council. Specifically, there were power struggles between the leaders of the different groups that signed the 1997 CPA. Though the five groups came together to form the United Democratic Salvation Force (UDSF) at the time of the agreement, the UDSF never actually operated as a coherent group. Rather, Riek Machar (leader of the SSIM) became the clear leader of the UDSF as the president of the Coordinating Council, and the leaders of the other four groups fought for positions in the Council and in the GoS. In addition to the intergroup fighting, several of Machar’s men left the SSIM to form their own groups in the hopes of gaining power and positions in other parts of the South or in Khartoum. It did not help matters that Sudan’s President Omar al-Bashir actively supported the various southern groups as they struggled for power, supplying some with weapons and equipment.
While the UDSF and Coordinating Council slowly unraveled, the GoS began a military campaign against the SPLA, and the SPLA continued its military campaign against the GoS. In other words, the situation on the ground did not change substantially after 1997. The new millennium, however, brought several changes within the southern rebel groups, among the groups, within the GoS, and in the international community. By the beginning of 2000 it was clear to those inside and outside of the South that the 1997 CPA was not being implemented by the GoS. It came as no surprise, then, when Riek Machar stepped down from the UDSF and cut all ties with Khartoum. Though al-Bashir appointed a replacement, the UDSF ceased to function with Machar’s departure. At the same time that this collapse took place, the GoS set about repairing its hostile relationships with its neighbors, and while it was doing so an event on the other side of the world would instantly change Sudan’s status in the international community: the September 11 terrorist attacks. Sudan had been on the U.S. “State Sponsors of Terrorism” list since 1993. Though there were several reasons for this designation, the main reason was that Sudan allowed Osama bin-Laden to become a resident after he was expelled from Saudi Arabia in 1992. Following negotiations with the U.S. and bin-Laden himself, he left Sudan in 1996. Despite bin-Laden’s departure, members of al-Qaeda remained in Sudan.
Five years later, when it became clear that al-Qaeda was responsible for the September 11 attacks, Sudan began working with the U.S. and the international community to find and capture terrorists inside its borders. This new cooperation was a complete reversal in Sudan–U.S. relations. In the year following this shift in Sudan’s status in the international community, two critical events changed the course of the war: there was a reconciliation between Riek Machar and John Garang, and the first in-person meeting took place between John Garang and Omar al-Bashir. Though Machar left the USDF and cut ties with the GoS in early 2000, he still operated a rebel group that had splintered from the SPLA. In the period following the 1997 CPA, several groups splintered from Machar’s USDF (and his original SSIM), which led to a situation in which there were multiple rebel groups fighting one another and the government. Following Machar’s departure from the USDF, interfactional fighting decreased dramatically. There were several agreements between leaders of SPLA splinter groups that reunited these groups with the original SPLA. In January 2002 Riek Machar and John Garang signed a formal agreement to reconcile and work together toward their common goals: peace and self-determination for southern Sudan. With these groups now united as the SPLA, Garang explained, a real peace process could begin with the GoS (quoted in Cobb Reference Cobb2002; IRIN 2002).
In addition to Machar and Garang’s January 2002 meeting, another critical meeting took place just seven months later: Omar al-Bashir and John Garang met in person for peace talks. This face-to-face meeting in July—the first time the Sudanese president and leader of the SPLA actually met—was the beginning of talks that led to the Machakos Protocol (Government of Sudan 2002) in July and a comprehensive truce in October 2002. The Machakos Protocol, considered by many to be the first major step in the peace process that led to the 2005 CPA, explicitly stated both sides’ willingness to begin a peace process that would address the South’s right to self-determination, religious freedom, participation in the GoS, and other concerns. The groundwork laid in the Protocol gave the parties to the conflict and involved third-party actors a framework within which to reach the 2005 CPA, which provided the details for the implementation of all issues mentioned in the Protocol. The comprehensive truce, for its part, was designed to halt violent conflict while the GoS and SPLA continued peace talks. It was designed to give all actors in the negotiations the space within which to work without fear of a return to full-scale war. Though there were acts of violence by both sides after the truce, large-scale battles were minimal. In 2003 and 2004 peace talks continued, which resulted in an agreement between the GoS and the SPLA to allow international observers to monitor the comprehensive truce as well as an agreement to share oil wealth.
Three years after the comprehensive truce, John Garang and then First Vice President of Sudan Ali Osman Mohamed Taha signed the 2005 Comprehensive Peace Agreement in the presence of several international mediators (Government of Sudan 2005). The 2005 CPA addressed most of the concerns of both sides, including those covered in the 1997 CPA. The major difference between the two CPAs is that the latter provided specific details for the implementation of the agreement and mechanisms for monitoring its implementation. For this reason, the 2005 CPA garnered much more support among the people of southern Sudan and in the international community as a whole. Equally important, this agreement, unlike the 1997 CPA, was negotiated and signed by the SPLA, which represented the vast majority of southern rebels by January 2005.
The 2005 CPA addressed many issues, including self-determination for the South, the role of Southerners in Sudan’s national government, the creation of the Government of South Sudan (GoSS), and the division of oil revenue (see Toft Reference Toft2010). Regarding self-determination, the 2005 CPA contained the same provision for a referendum on secession as the 1997 CPA, but the new agreement included a specific six-year timetable in which the referendum would take place, and details on how the referendum would happen and who would monitor it. Though there were many doubts about whether the referendum would actually happen, in January 2011 the referendum took place, and southern Sudanese citizens voted more than 98 percent in favor of secession (BBC 2011). On July 9, 2011, South Sudan’s goal of self-determination was realized when it became the world’s newest independent country.
One major difference between the 1997 and 2005 CPAs was the latter’s creation of a specific power-sharing agreement. Whereas the first CPA acknowledged Southerners’ grievances about not being included in Sudan’s national government but did little to change the situation, the second CPA completely reorganized the executive branch of the GoS, changing the position of first vice president to one that could only be held by a Southerner. This position, to which Garang was appointed, had key powers, including the appointment of various officials in the national government and the responsibility to protect the rights of Southerners living in the North. Along with reorganizing the executive branch, the CPA reconfigured Sudan’s legislature, the National Assembly, setting aside 30 percent of the four hundred and fifty seats to Southerners. The changes to these two branches of Sudan’s national government gave Southerners much more access to the political system than they had previously experienced (see Toft Reference Toft2010).
In addition to the new power-sharing arrangement, the 2005 CPA created the Government of Southern Sudan (GoSS). The agreement also included details of how the GoSS would operate, stipulating that it would have its own autonomous legislative, executive, and judicial branches to govern all states in southern Sudan. The most pressing tasks of the legislature, the Assembly of Southern Sudan, were to draft and approve the Constitution of Southern Sudan and hold elections for seats in the legislature. For the executive branch, the president of Southern Sudan (John Garang), who would also become the first vice president of Sudan, had the power to appoint an Executive Council of Ministers and establish any institutions deemed necessary. Along with the legislative and executive branches, the CPA called for the creation of an independent judiciary, to include a Supreme Court of Southern Sudan, a Court of Appeals, and any other courts listed in the new constitution. This final branch of government provided Southerners with a legal system that would not be based on Shari’a law, one of the major concerns of Southerners that had not been adequately addressed in previous agreements.
The final issue that was resolved by the 2005 CPA was an especially important one for the interim period leading up to the referendum, but it also had major implications for the future South Sudanese state: how Sudan and South Sudan would share revenue from oil exports. Earlier agreements had not dealt with the problem of oil-wealth distribution, so this was new ground for both parties. The CPA divided revenue from oil by giving 2 percent to the local areas surrounding the oil fields, 49 percent to the GoS, and 49 percent to the GoSS. In addition to the specifics of wealth sharing, the CPA created the National Petroleum Commission, chaired by the presidents of the GoS and GoSS, to monitor oil production and revenue. The Commission’s goal was to ensure fair reporting and distribution of profits according to the provisions of the CPA. Though there were a few important issues left unresolved (e.g., whether Abyei would be part of the North or South), the 2005 peace agreement covered many intractable issues and garnered the necessary support of the GoS and the SPLA.
Hypothesis 1: Rebel Group Grievances
Many of the changes that led to the 2005 CPA began in 2002 in the months leading up to the war’s last comprehensive truce in October. During this year there were two changes in the SPLA’s grievances against the government, both of which involved proposals by the U.S. The first change came when the U.S. offered its support of Sudan’s unity (i.e., no independence for South Sudan). Though the Clinton administration had been involved in the conflict by indirectly providing resources for the SPLA in the mid-1990s, in 2001 the Bush administration made Sudan a priority. Even before the need to work with Sudan on the “War on Terror” revealed itself, oil companies and others in the U.S. were lobbying Bush to remove the sanctions on companies doing business in Sudan from the June 2001 Sudan Peace Act. In August 2001 the administration publically announced that it did not approve of the sanctions. After the terrorist attacks one month later, the administration made it clear that it planned to work with Sudan’s government in joint counterterrorist efforts (Huliaris Reference Huliaris2006; Johnson Reference Johnson2011).
The new “War on Terror” and the old need for access to additional sources of oil influenced the new interest on the part of the U.S. in a peaceful resolution of the decades-old civil war in Sudan. In January 2002 the U.S. demonstrated its good will toward Sudan by announcing that it supported an agreement between the warring parties that would maintain Sudan’s unity. This marked a minor change from the previous U.S. position that had no specific discussion of self-determination. In terms of the SPLA’s struggle for autonomy, this shift in U.S. foreign policy and the new alignment with the GoS led to an increase in SPLA grievances because the U.S. opposed autonomy or independence for South Sudan. This was not a major increase in the SPLA’s autonomy grievance because this new support from the U.S. did not translate into a concrete change in Sudan’s policy.
Three months after this shift in the U.S. position, however, it made a proposal that led to a decrease in SPLA grievances. Once the U.S. became involved in negotiations between the North and South, it began making proposals in an attempt to keep the dialogue between the two sides going. In April 2002 the U.S. suggested that the GoS and SPLA share revenue from oil exports. In addition, the U.S. pushed for freedom of religion throughout the country. Though the GoS did not agree to either, this marked the first time that an influential third party called for wealth sharing. The fact that the South now had another actor trying to persuade Sudan to share revenue and allow freedom of religion increased the probability that these changes actually would take place. This new hope in future oil-revenue sharing and exemption from Shari’a law decreased the SPLA’s grievances against the government. Three months after the U.S. made its suggestions, the SPLA and the GoS signed an agreement on religious freedom, which dramatically reduced SPLA grievances. From the July 2002 peace agreement to the 2005 CPA, the SPLA only carried out two more reported acts of violence, showing their willingness to participate in rather than spoil the peace process.
Hypotheses 2 and 3: Rebel Group Cost of Using Violence
In the time leading up to the 2005 CPA, the improved relations between Sudan and its neighbors meant that the SPLA experienced a major decrease in its capacity to use violence. In December 1999 Sudan began repairing its relations with its neighbors, beginning with the resumption of diplomatic relations with Uganda. This was followed with a renewal of diplomatic ties with Eritrea and Ethiopia in January and March 2000, respectively. With these renewed diplomatic relations, each country expressly agreed to stop providing arms, equipment, sanctuary, and sometimes troops to rebels in other countries. In addition, Sudan made agreements with both Ethiopia and Uganda to work together to oust rebels operating along the borders. These improved relations with Sudan meant that the SPLA lost the support of important allies, which included use of land for SPLA rear bases, money, weapons, and equipment. Because these countries had also been used as go-betweens for U.S. weapons and equipment for the SPLA, the loss of support from these countries was a major blow to the SPLA’s access to the resources required to carry out violence.
The SPLA’s loss of support from Sudan’s neighbors did not automatically mean a loss of support from the U.S. All of this changed, however, the next year with George Bush taking office and the September 11 terrorist attacks. Bush made access to Sudanese oil a priority upon taking office, but this could only be accomplished with better relations between the two countries. After September 11 the priority shifted to cooperation with Sudan on counterterrorism. These developments meant a loss of U.S. support for the SPLA, which was a major change from both the Clinton administration’s position during the 1990s and the U.S. Senate’s position when it passed the Sudan Peace Act in July 2001. The Act had called for sanctions against Sudan and companies doing business with Sudan, condemned human rights abuses, and provided funding for rebel groups operating in southern Sudan. After September 11 the loss of resources from the U.S., along with the loss from neighboring countries, led to a major decrease in the SPLA’s capacity to use violence, and the SPLA’s capturing of towns decreased dramatically.
The years leading up to the 2005 CPA were also marked by GoS military campaigns, aerial attacks, and external pressure for a peaceful end to the decades-long war. Between 2000 and the 2005 CPA, the GoS carried out nine separate military offenses, including aerial bombings and ground attacks by the army and government-sponsored militias. Several of these offensives lasted for over a month, resulting in dozens of casualties and even more injuries. The three months leading up to the June 2002 peace talks saw a “major offensive” on the part of the GoS in two southern Sudanese regions, a “dramatic increase” in aerial bombings, and the retaking of three towns that had been controlled by the SPLA. In fact, from the beginning of 2002 to the signing of the Machakos Protocol in July, thirty-six civilians were reportedly killed, including foreign NGO staff members. Unlike the period before the 1997 CPA, there was no doubt that the GoS was willing and able not only to retaliate against SPLA violence, but also to go on the offense militarily.
In addition to Khartoum’s increasing military success, both the SPLA and the GoS were getting pressure from other states to end the civil war peacefully. With the U.S. and others interested in Sudan’s increasing oil output, both sides in the civil war were urged to begin the peace process anew, especially since a great deal of the fighting occurred near the oil fields. These regions were (and continue to be) marked by insecurity and threats against foreign oil workers, which hurt Sudan’s productivity in extracting and exporting oil. An end to the civil war would also help the GoS focus on counterterrorism, and Sudan’s status as a safe haven for terrorists was a priority in the U.S. “War on Terror.” As mentioned above, Clinton’s policy had been to indirectly supply help to the SPLA and others fighting against Sudan’s regime. It was obvious, however, that this strategy was no longer possible with the need for new sources of oil and counterterrorism allies.
Instead, the Bush administration sent a special envoy in an attempt to bring the two sides to the negotiating table. Washington’s new relationship with the GoS meant that it supported peace, but on Khartoum’s terms (no independence for the South). This stance was contrary to SPLA interests, causing an increase in grievances. At the same time, it showed U.S. willingness to be actively involved in brokering, monitoring, and implementing the terms of a peace deal, giving the GoS justification for any acts of retaliation against SPLA attempts to spoil the peace process, and possibly U.S. assistance to quell any such attempts. In addition, a January 2002 ceasefire between the GoS and the SPLA for the Nuba Mountains region allowed international monitors to ensure that both sides were adhering to the agreement (IRIN 2002). This shift in U.S. policy and the international community’s (new-found) commitment to end the war, therefore, posed an increased threat of retaliation for violence, making any SPLA attempts to spoil the peace process more costly.
Hypothesis 4: Rebel Group Benefit of Using Violence
After the 1997 CPA, the SPLA and its northern allies in the National Democratic Alliance (NDA) continued capturing towns and providing government-like administrative services to the residents in these territories. Unlike the period before the 1997 CPA, however, the GoS also began recapturing more and more towns, which decreased the SPLA’s likelihood of military victory. In the late 1990s and early 2000s the GoS began retaking towns and villages, which meant that the SPLA was being militarily defeated with increasing frequency. Beginning in 2000 the GoS began retaking key areas and transportation routes, including parts of Kassala, Bahr al-Ghazal, Upper Nile, and the railway leading into Wau. During this period, the SPLA was able to capture towns as well, so what emerged was a relatively equal military balance of power between the SPLA (and NDA) and the GoS. Both sides held large swathes of land, with each side adding to and losing land in the process. Though the SPLA was still able to hold and gain territory, its potential for a clear military victory decreased because of an increase in the number of military defeats and increasing loss of territory in the years leading up to the Machakos Protocol.
Hypothesis 5: Rebel Group Position in the Peace Process
Unlike the period leading up to the 1997 CPA, the SPLA played an active role in the peace process leading to the 2005 CPA. In January 2002 the GoS and the SPLA agreed to a ceasefire in the Nuba Mountains. Though this ceasefire only covered a portion of the war zone, it was an important shift for both the GoS and the SPLA, which had been carrying out large-scale military attacks against each other since the 1997 CPA. Six months after the ceasefire, both parties signed the landmark Machakos Protocol, and Garang and al-Bashir met for the first time. In September 2002 they expanded the January ceasefire to include all of southern Sudan, and in October they signed the comprehensive truce to halt all military operations during the peace process. As previously mentioned, both sides still carried out acts of aggression during this period, but the large-scale violence of the late-1990s and early-2000s was over. Throughout the period between the comprehensive truce and the January 2005 CPA, several more agreements were signed between the SPLA and the GoS, including an agreement on oil-wealth sharing in January 2004. By 2005, then, it was clear that the SPLA was an active participant in the peace process, and as such would have lost the material, political, and reputation benefits that came from being in this role had they decided to spoil the process.
Comparing the Peace Processes and Exploring Alternative Explanations
After examining the SPLA’s decisions during the peace processes leading up to the 1997 and 2005 CPAs, we also need to compare the changes in the SPLA’s grievances and cost–benefit analysis during both processes. First, SPLA grievances increased before and as a result of the 1997 CPA, while their grievances decreased in the period leading to the 2005 CPA. Second, the SPLA’s capacity to use violence increased dramatically leading up to the first CPA, but its capacity decreased before the second CPA. In addition, while the threat of retaliation against the SPLA for its use of violence did not change much before the 1997 agreement, this threat increased before the 2005 agreement, due in part to the (newly) watchful eyes of the international community. There was also a clear shift between the SPLA’s high potential for military victory in the period before the 1997 CPA and its decreasing potential during the peace process leading to the 2005 CPA. Finally, the SPLA’s position outside of the 1997 peace process increased the potential benefits of spoiling it, while its active participation in the 2005 peace process increased the costs of acting as spoilers, namely the potential loss of political power and material goods. These findings support the hypotheses, and together provide strong support for the argument that rebel groups’ grievances and cost–benefit analyses influence their decisions to spoil or participate in a peace process.
Though the evidence supports the hypotheses, it is important to identify whether it can also be used to support alternative explanations of leaders’ differing decisions of whether or not to spoil the 1997 and 2005 CPAs. One of the most widely cited explanations for actors’ decisions during conflict is greed, especially in areas like Sudan that are rich with natural resources. Arguments about the role of greed maintain that leaders’ decisions are motivated by the desire to control the access to and profits from oil, mineral deposits, and other primary commodities.Footnote 9 If rebel leaders’ greed or material covetousness were able to explain SPLA leaders’ decisions during the 1997 and 2005 peace processes, we would have seen similar decisions each time because the material wealth of the SPLA, or its potential for wealth, did not change between the two agreements. Specifically, neither the amount of oil in Sudan, nor the SPLA’s access to this oil, changed after the 1997 peace process. Regarding access, both processes led to agreements that included wealth sharing, including the particular sources of revenue to be shared and the formula for revenue allocation. Though the 2005 CPA included a more robust plan for wealth sharing, the 1997 agreement would have given the SPLA much more revenue (as part of Sudan’s federal system) than it had before the agreement, including a portion of state and federal taxes, oil revenue, and control of development aid. If material considerations were the only, or even the primary, motivation for the SPLA, there would have been no reason to spoil the 1997 peace process and resulting agreement.
Another common explanation for actors’ decisions before and during a conflict is based on the importance of ethnicity. Many ethnicity-based arguments focus on ethnic fractionalization, identifying the direct or indirect role of the number of ethnic groups in a country in determining the likelihood that actors will use violence.Footnote 10 Most ethnic fractionalization arguments hold that greater fractionalization will increase the likelihood that actors will use violence for various reasons, including historical hatred between groups or ethnically based competition for scarce resources. If ethnic fractionalization were able to explain SPLA leaders’ decisions in 1997 and 2005, we should have seen similar decisions during both peace processes because the number of ethnic groups in the region did not change in the time between the agreements. To be sure, ethnicity has played a role in SPLA decisions, the formation of SPLA splinter groups, and fighting between the SPLA and other southern Sudanese groups during the war. But there is no evidence to support an argument that ethnic fractionalization affected SPLA leaders’ specific decisions to spoil or participate in the two peace processes.
Conclusion
Exploring the divergent decisions of the SPLA and other rebel groups in southern Sudan sheds light on the fact that not only are rebel groups not unitary actors, but they often experience internal, and sometimes violent, conflict. Examining the SPLA’s changing decisions over time, however, shows that as internal and external factors changed, the same group that spoiled one peace process became a key participant in the next. After tracing the processes that led to the 1997 and 2005 CPAs, some important difference begin to stand out.
The more comprehensive theoretical framework used in this article allows researchers to explore multiple, often-changing variables, which account not only for actors’ decisions at one point during a peace process, but also changes in their decisions over time. Understanding these changes over time, in turn, can highlight the reasons that a group may participate in one peace process and spoil another. In addition to understanding these decisions, it is important that researchers examine rebel groups not as unitary actors, but as multiple actors with often conflicting interests and strategies. This article focused on the original SPLA, but it included discussions of splinter groups, which acted separately. A vital direction for future research is studying the role that splinter groups play in peace processes. Empirically, there is a correlation between the presence of multiple rebel groups and failed peace processes, given that all of the current armed conflicts in Africa with failed or stalled peace processes also have more than one rebel group opposing the state. Continuing research on rebel groups’ changing decisions over time, and the impact of the presence of multiple rebel groups, will provide scholars and policymakers with more nuanced approaches to getting key rebel groups to the negotiating table, and keeping them there. As the findings from this and other articles have shown, it is better to get as many actors as possible involved in a peace process, because the costs of spoiling the process increase and the benefits of spoiling decrease once a group has a seat at the table.
Acknowledgments
The research for this article was made possible by grants from Ramapo College of New Jersey and the Ramapo College Foundation. I would like to thank my research assistant, Zoie G. Schachter, and the managing editor and anonymous reviewers from the African Studies Review for their insightful suggestions.
