PREVIOUS RESEARCH AND THEORETICAL APPROACH

Two research trends have served to further our understanding of the challenges that peaceful politics face in war-torn countries. A first strand highlights the dilemmas encountered in war-to-peace transitions and focuses on the structural features of societies shattered by armed conflict. For instance, institutions and party structures are weak, civil society is generally politicised or has minimal influence, spoilers set on ruining the peace process are present, infrastructure is poor and there is dependence on large-scale international involvement (Paris Reference Paris2004; Lyons Reference Lyons2005; Jarstad & Sisk Reference Jarstad and Sisk2008; Brancanti & Snyder Reference Brancanti and Snyder2013).

These conditions make the advent of multi-party politics and the institutions sustaining democracy precarious. Moreover, war-to-peace transitions are often accompanied by state-building processes, emphasising the necessity to develop capable and legitimate state institutions to underpin the conduct of democratic politics. While a necessary step, this introduces further difficult conditions since institution-building in essence requires not only fundamental reform, but also time (Paris & Sisk Reference Paris and Sisk2009). Taken together, this research has pointed to the conflict-inducing aspects of elections, especially in times of transition.

A second strand focuses more directly on violent electoral competition and views violence as a strategy parallel to other constitutional or non-constitutional strategies, which governments and opposition parties use to further their influence during electoral periods. Election-related violence has been studied both as a strategic form of political violence used by elites and as more grass-root driven processes, including ethnic rioting and violent protest.

Part of this research stresses the features of the electoral process, in particular the incentive structures shaped by institutional weakness and constitutional constraints, the electoral system, electoral management bodies and the nature of international involvement (Hafner-Burton et al. Reference Hafner-Burton, Hyde and Jablonski2010; Daxecker Reference Daxecker2012; Fjelde & Höglund Reference Fjelde and HöglundForthcoming; Arriola & Johnson, ms). For instance, Wilkinson's (Reference Wilkinson2004) study on ethnic riots in India suggests that electoral incentives at the town level explain where Muslim–Hindu riots break out, whereas state-level electoral motivations decide if the police will be called in to prevent riots.

Other studies emphasise more general features of society, such as the combination of social cleavages and patrimonialism. Patrimonialism is a feature of societies which increases the stakes of elections, especially when intertwined with societal features which price ethnicity as the most important political asset (Arriola Reference Arriola2009). The importance of patrimonial politics for electoral conflicts can be witnessed, for example, in Kenya, Zimbabwe and Côte d'Ivoire, where local land grievances are used by opportunistic politicians for their own political purposes (Boone Reference Boone2012; Boone & Kriger Reference Boone, Kriger and Bekoe2012). Conflicts between Christians and Muslims in Indonesia have been based on patronage networks and clientelism, where the triggering factor was the upcoming election (van Klinken Reference van Klinken2001). These studies point to the interlinkages between the formal political arena and the informal structures which are used by political elites to gain and maintain power.

This article adopts a novel approach to studying the violence-election nexus in South Sudan. To the case of South Sudan, we apply the theoretical framework developed by Höglund et al. (Reference Höglund, Jarstad and Söderberg Kovacs2009), which suggests three sets of factors as important features in explaining the conflict-generating dynamics of elections: stakes, actors and institutions. While other factors also come into play, such as ethnic and socioeconomic configurations in the country, resource scarcity or path dependency, the theoretical framework is useful to capture the dynamics in South Sudan, since its constituent components include factors of relevance both from a peace-building and state-building perspective.

Stakes highlight the price attached to winning elections. The likelihood of elections becoming violent is premised on the importance of the elections. In transitioning countries, the stakes of elections are generally higher than in consolidated democracies. Access to state power becomes a means to secure personal wealth and provide significant benefits for one's own constituency (Bratton & van de Walle Reference Bratton and van de Walle1997; Posner Reference Posner2005). Presidential and governmental power is of particular importance, especially in highly centralised states, since control of the state's economic resources is exercised at the centre. In non-democratic countries, state assets can also be used to oppress political opponents to gain an advantage during times of election. For instance, the police and military may be manipulated to make extra-legal arrests of regime critics, break up political meetings of opposition parties, or intimidate opposition voters and candidates.

Actors point to a component that deals with the political players involved in the electoral process. Political parties are considered the key agents in electoral processes, because they establish the link between the voter and the institutions of democracy. In countries emerging from violent conflict, the main political actors are generally strongly militarised. While non-militant political parties may exist, many of the electoral competitors may be formerly armed actors who only recently became formal political parties. In addition, in many developing countries the political parties are fairly weakly institutionalised and have insufficient resources (Panebianco Reference Panebianco1988; Diamond Reference Diamond1999; Randall & Svåsand Reference Randall and Svåsand2002). This holds true in many war-torn societies as well. The main challenge from the perspective of violence, is the involvement of militants in politics. Democratic, non-violent behaviour and norms take time to develop. The risk is considerable that violence will be used as a means to repress opponents, or, if faced with electoral loss or political marginalisation, take up arms (Lyons Reference Lyons2005; Manning Reference Manning2007; Söderberg Kovacs Reference Söderberg Kovacs, Jarstad and Sisk2008).

Institutions emphasise the structures of the state in which elections take place. The institutional framework – including the electoral system, the electoral law, electoral management bodies (EMBs), police and judiciary ‒ determines the rules of the game in electoral politics. A key issue to consider with regard to institutions and conflict management in a post-war context is the extent to which the institutions increase polarisation by excluding important actors with the capacity to foment violence. The representativeness of institutions such as the executive, parliament and local government is, for instance, influenced by the type of electoral system. Majoritarian systems may to a larger extent than proportional ones encourage winner-takes-all dynamics and create permanent minorities, which may have a violence-inducing influence on electoral politics (e.g. Sisk Reference Sisk1996; Reynolds Reference Reynolds1999; Lindberg Reference Lindberg2005).Footnote ² The representativeness of key institutions, such as EMBs, the police and judiciary, is also fundamentally determined by appointment and recruitment processes, which may be based on political considerations and partisan interest, rather than merit. A second aspect of importance relates to the capability and independence of institutions to fulfil the roles they are meant to perform. For instance, electoral laws may be designed to prevent militant actors from standing in elections and may prohibit ‘hate speech’ that might encourage violent mobilisation. However, if the police and judiciary are weak or politicised and thereby unable to enforce the laws, these laws will have little influence on behaviour around elections.

We seek to trace the connection between election and violence in a broad sense. We do not confine our study to violence in the immediate proximity to elections, but include violence prior to and after the elections. We also analyse both communal and state-based armed conflicts, and the extent to which they are interwoven with electoral dynamics.Footnote ³ For this reason, we do not use the term electoral violence, which is considered a specific form of political violence in much of the academic literature (Höglund Reference Höglund2009; Daxecker Reference Daxecker2012). Communal conflicts constitute ‘conflicts between non-state groups that are organised along a shared communal identity’, while state-based conflict is a ‘conflict with a contested incompatibility that concerns government or territory … between two parties, of which at least one is the government of a state’ (UCDP 2014). In our study we focus primarily on the conflicts which originate in relationships between the elites, but throughout the analysis we highlight the interlinkages to local conflict dynamics (Kalyvas Reference Kalyvas2006). Moreover, the South Sudan case points to an additional factor with regard to institutions which is important to consider. In states completely dominated by one political party, the real political competition takes place within the party, rather than at the actual elections.

ELECTORAL POLITICS AND PATTERNS OF VIOLENCE IN SOUTH SUDAN

On 9 January 2005, Sudan's longstanding North–South war ended through the signing of the Comprehensive Peace Agreement (CPA) by the Government of Sudan and the rebel movement SPLM/A. The agreement ended Africa's longest conflict that killed more than 2 million people. The war had its origin in the centralisation of economic and political power in Khartoum, in the north of Sudan, whereas the peripheries – the south of Sudan in particular – remained severely marginalised. The North–South divide was further entrenched by the north being predominately Muslim and Arabic, whereas the south is predominantly Christian and African. The CPA fundamentally changed South Sudan as a political entity by granting the region extensive autonomy and by establishing the Government of South Sudan as the prime ruling authority of the area. The agreement also resulted in power-sharing institutions, such as the Government of National Unity with representation from the north and south, wealth-sharing arrangements for the oil, and the formation of new armed forces. On 9 January 2011, a referendum was held in which 99% of the electorate voted in favour of independence and South Sudan was born on 9 July 2011 (LeRiche & Arnold Reference LeRiche and Arnold2012).

THE STAKES: THE PRICE OF STATE POWER

The stakes involved in the political struggle in South Sudan are immense. The structure of the South Sudanese state – and the party governing it – is very hierarchical and puts the winner in this contest in a powerful position. South Sudan has a very strong presidential system in which the parliament only has limited power to counteract decisions by the president (Radon & Logan Reference Radon and Logan2014: 157). Power is also concentrated at the top level at the expense of more local governance structures.

The Government of South Sudan was established by the CPA and the Interim National Constitution of the Republic of South Sudan. However, state institutions are entirely dominated by the SPLM, in essence making it a one-party state (ICG 2014). The history of the SPLM is closely linked to John Garang, who led the organisation from its formation in 1983 until his death in 2005. Garang ruled the SPLM in a highly authoritarian manner, and during the war his leadership was frequently challenged. For instance, the SPLM split in 1991 after Riek Machar and Lam Akol accused Garang of dictatorial behaviour and formed a separate group. This led to fighting between the two factions, with widespread atrocities committed by both sides (Jok Reference Jok2007). To answer to the demand for reform, and to consolidate power, Garang used a strategy of co-optation. While this policy caused interludes in the fighting, it did not result in any real changes in the organisation (Mamdani Reference Mamdani2014). The authoritarian manner in which the SPLM has traditionally been ruled has continued after Garang's death. While Garang's successor, Salva Kiir, used a ‘big-tent’ policy, which – at least on paper – incorporated various factions within the armed forces, it did not constitute any real institutional changes (ICG 2014).

The leader of the SPLM/A is set to preside over a sovereign government, and the territory it controls has significant oil resources and other important economic assets, including land. In rural South Sudan, control over land is economically vital for agriculturalists and pastoralists alike, and politicians in power have extensive power to determine landownership. The stakes are augmented, as all subterranean natural resources belong to the central government (Mertenskoetter & Luak Reference Mertenskoetter and Luak2012). Furthermore, corruption is widespread and the governance system has taken the form of a kleptocracy. Thus, political elites have extensive opportunities to enrich themselves (de Waal Reference de Waal2014). This dynamic is largely a continuation of the predation that took place during the North–South war when commanders in the South carried out extensive predation on the areas they controlled (Pinaud Reference Pinaud2014). The incentive to capture influential positions is further amplified, as the prospects of getting rich outside the government structures are very limited. Government appointments are highly prized in South Sudan because they represent one of the few opportunities to safeguard income and influence (ICG 2009). The institutionalised corruption creates further problems for those outside the government. Without connections within the government structures, individuals can face significant problems with administrative practicalities such as obtaining permits (UNMIS representative 2009 Int.).

Obtaining an influential position is also a way for elites to secure resources for their constituencies, which in the South Sudanese context generally means community, ethnic group or family. The peace dividends resulting from the 2005 peace agreement are limited and unevenly distributed. For many residents in South Sudan, living conditions have not improved. In fact, the agreement resulted in renewed suffering for some residents since several relief organisations reduced their contributions after the CPA. Some economic and infrastructure development has taken place in Juba, but other areas have seen no progress. The political and economic centrality of Juba means that decisions in the capital determine where development projects are located (Ibrahim Adam Mudawi 2009 Int.). Thus, the stakes at play in the political game in South Sudan are very high, which increases the risk for violence around elections.

ACTORS: THE IMPORTANCE OF WAR-LORD POLITICS

South Sudan gained its autonomy, followed by independence, from the 2005 peace agreement. The accord gave almost all power to the two main belligerents in the civil war, the SPLM and the Sudanese government, whereas other political movements were excluded (Mamdani Reference Mamdani2014). In post-agreement South Sudan, this situation endures and the political life is dominated by individuals who held positions within the rebel groups. Government positions are predominantly occupied by militaries. Moreover, influential generals are sometimes more powerful than ministers, dictating how ministers allocate the resources they have at their disposal (de Waal Reference de Waal2014). However, there are some prominent exceptions. For instance, the well-known politician Francis Deng, who in 2012 was appointed South Sudan's ambassador to the United Nations, has a civilian background (Miamingi Reference Miamingi2012). There are two dimensions related to the North–South war which have caused political discord. First, there are tensions between those who fought for independence and those who sided with the Sudanese government during periods of the North–South war. These conflicts originate in splits in the rebel movement during the war, which caused fighting between South Sudanese rebel groups. Second, tensions exist between those who remained in the country and fought in the war and those who fled and have returned after it ended. In general, returnees are more educated and believe they should take a leading role in politics. However, those who fought the war, also feel entitled to influential positions (e-mail conversation Tubiana 2015).

From the perspective of violence, the post-agreement political structures of South Sudan have made the threat of violence a powerful tool for gaining concessions and influence (LeRiche Reference LeRiche2014). In particular, elites are given prominent positions depending on how large a threat they constitute (SPLA veteran 2011 Int.). When the government of South Sudan was installed it was challenged by several armed groups. Many of these groups had fought against the SPLM during the North–South war, as Khartoum persistently used a divide-and-conquer tactic that resulted in extensive fighting between Southern factions (Johnson Reference Johnson2012). To deter the threat of these groups, the regime incorporated them in the governmental structure through a ‘big-tent’ policy, which meant that instead of fighting off rivals they were included in the state structures. The rank-and-file of these groups were included in the armed forces, while the leaders gained positions in the army and were sometimes included in the government. The most powerful of these groups was the South Sudan Defence Forces (SSDF), which were incorporated through the Juba Declaration signed in January 2006 (Young Reference Young2006). In the following years, co-optation continued through a widely used policy of granting amnesties to different factions that fought against the regime (ICG 2014). Although this policy ended several insurgencies, it also created several problems.

First, it encouraged armed rebellions, since it meant that military leaders could defect, fight for a while, be granted amnesty and then be reintegrated into the governmental structures, often with a higher grade and salary. Many of the leaders were given high positions within the army, creating strong incentives for military commanders to launch rebellions. Such dynamics are illustrated in the case of Peter Gadet, who has defected twice (in 2011 and 2013) since the signing of the 2005 peace agreement (BBC 11.7.2014). Moreover, individuals have strong incentives to join rebellions, since it is beneficial to be part of the army. The lowest-ranking soldiers in the army are paid about US$140 a month, which is a high salary considering that most of the population lives on less than US$1 a day (Mamdani Reference Mamdani2014). Thus, the opportunity costs of joining a rebellion are low.

Second, the policy of co-optation is very expensive for the government to maintain. Prior to the rebellion in mid-December 2013, it was estimated that 50% of the government budget was used to pay the armed forces (Mamdani Reference Mamdani2014). When the peace agreement was signed, about 200,000 Southerners were part of the Sudanese army and ‘other armed groups’, primarily militias that had fought on Khartoum's side during the war. Five years later this number was reduced by almost 90%, while at the same time the army numbers increased from about 50,000 to more than 300,000 (de Waal Reference de Waal2014). As most fighters in these militias were Nuer, it also created an army where a majority came from this community (ICG 2014).

Third, this policy has created a deeply divided army (ICG 2014). Reconciliation between different factions has been hampered, and to a large extent the SPLA has become a coalition of ethnic militias rather than a national army (Mamdani Reference Mamdani2014). The deep division within the SPLA was a key factor in explaining the rapidity with which events unfolded in late 2013.

Fourth, local grievances and ethnic identities have been manipulated by local and central elites in their struggle for political power, thereby linking tribal politics to politics at the national level (da Costa Reference da Costa2012). Typically, political players in Juba instigate communal conflicts between communities to sustain their own significance. For instance, high-ranking officers within the army, as well as important politicians, often provide weapons to their own community to ensure political support. This practice has become a prevalent feature of politically tense periods, such as elections (James Ninrew 2011 Int.). In addition, local politicians use similar tactics in order to empower themselves (Sørbø Reference Sørbø2010; International academic 2011 Int.). This type of political positioning has been particularly fierce in Jonglei, which has also been the state most struck by violent conflicts (Young Reference Young2010).

Finally, the long history of war in the region has made retaliatory dynamics crucial, and many conflicts are retributions for earlier conflicts (Harragin Reference Harragin2011). These dynamics make it easier for military ‘strong men’ to gain prestige. High-ranking officers within the army, as well as important politicians, often provide weapons to their community, which ensures that they remain in a powerful position (James Ninrew 2011 Int.).

To conclude, the political landscape in South Sudan is dominated by actors associated with the previous civil war. In this milieu, political power struggles often become armed, and political entrepreneurs frequently use weapons in order to empower themselves. In the following, we will turn to examining the institutions in which these actors operate.

INSTITUTIONS: FEEBLE STATE AND PARTY INSTITUTIONS

When South Sudan gained autonomy it started off from a deplorable situation. Not only was the area shattered by decades of war – and animosities among actors and communities were high – but state structures and infrastructure were also extremely weak, or non-existent. In fact, practical matters such as offices for the ministers were key problems that occupied the regime and international donors during the first years after the signing of the CPA (Senior international observer 2009 Int.). A main issue with regard to the institutions is that the regime in Juba has completely failed to form independent institutions with broad representation (ICG 2014). While co-option has been used as a way to circumvent threats from the opposition, inclusion in state institutions has been dependent on the actors’ military capability, rather than its representativeness. Moreover, the institutions are dependent on the SPLA, which has dominated most important aspects of life in South Sudan, including the government (de Waal Reference de Waal2014). After SPLA, the churches are the institution which reaches most people in South Sudan and they often play an important role in attempts to solve inter-communal conflicts (Archbishop Deng 2011 Int.; ICG 2014). In addition, the church has occasionally challenged the government. For example, the archbishop of Juba, Paulino Lukudu Loro, has criticised the widespread corruption and Dinka dominance in the government (National Catholic Reporter 22.10.2012).

Institutional progress since independence has been piecemeal, and the dysfunction of political institutions continues to be a fundamental problem. It shapes conflicts in the country in several ways (de Waal Reference de Waal2014). First, a crucial institutional problem in South Sudan is the lack of established, strong democratic political parties. This is particularly troublesome, given the strong standing of the president, ‘where the powers concentrated in the presidency under the 2011 Transitional Constitution are even greater than those normally vested under presidential systems’ (Radon & Logan Reference Radon and Logan2014: 156). An important factor for the 2013 crisis to unfold was that President Kiir cancelled several important SPLM meetings – a move that he was able to make because of weakly institutional structures of the SPLM. In particular, the party lacks a democratic process for leadership change. The opportunities to challenge incumbent leaders are restricted because open voting, rather than secret ballot, is the standard voting procedure. There has been opposition to open voting, which may discourage votes against the top leaders – President Kiir in particular. In fact, this issue was critical in the political contestation resulting in the violence that erupted in December 2013 (ICG 2014). Nevertheless, the power of the President is not completely uncircumscribed and presidential decisions are sometimes blocked. One well-known example is how the general assembly of South Sudan in August 2013 stopped the President's nomination of Telar Deng to become the next minister of justice (Sudan Tribune 13.8.2013). South Sudan also lacks established opposition parties. In the 2010 elections, most opposition politicians ran as independent candidates. This meant that elections stood between influential individuals rather than political parties, which increased conflict-generating factors such as patrimonialism and ethnic favouritism, especially since the politicians involved were military men.

A second problem related to institutions and elections concerns the electoral management bodies. The National Election Commission (NEC) was assigned to administer the 2010 elections. This institution attempted to increase the credibility of the elections, but the absence of voter training and widespread irregularities related to the polling, counting and tabulation, undermined the legitimacy of the elections (Carter Center 2010). The NEC is also tasked with administering the next elections, but the delay of the elections, repeated flaws in the electoral preparations, such as rescheduling of important meetings within the SPLM, and the chaos that has prevailed since late 2013, has severely circumvented the ability of this institution to fulfil its task.

Third, there is no functional criminal justice system in place, and there are several parallel judicial systems operating – customary as well as statuary – and they are not sufficiently coordinated. In addition, the lack of trained civil servants causes problems for the police. In this context, arbitrary detentions and other human rights violations are common (HRW 2011). A further problem for the citizens of South Sudan is that the regime in Juba tends to act in a biased manner – often favouring Dinka over non-Dinka – and the risk of facing sanctions from the state varies depending on the communal affiliation of the perpetrator (Brosché Reference Brosché2014). This has seriously diluted state institutions and undermined the rule of law, which increases the risk of political competition turning violent (UNMISS 2014).

Fourth, local government institutions are also very weak. This raises the risk for violence – communal conflicts in particular – as it undercuts local conflict resolution mechanisms that can solve such conflicts before they become violent (Brosché Reference Brosché2014). The Government of South Sudan was established by the CPA and the Interim National Constitution of the Republic of South Sudan, which both emphasise decentralisation of power and local democracy. The SPLM also employs a rhetoric that focuses on local governance (Rolandsen Reference Rolandsen2005). However, such structures hardly exist in reality and the opportunities for local establishments to influence rules, or to control their execution, have been severely restricted (Rolandsen Reference Rolandsen2007). In fact, in many areas of South Sudan no functional, legitimate, local governing institutions exist (Young Reference Young2010). The move away from decentralisation was further established in the 2011 Transitional Constitution, which strengthened the national government at the expense of local structures (Radon & Logan Reference Radon and Logan2014). How the centre influences local policies in South Sudan can be illustrated by the way in which a governor is chosen. In the first years after the signing of the CPA, governors were not elected but directly appointed by the president (Rolandsen Reference Rolandsen2007). It was not until 2010 that the first election of a governor took place, and this process is still primarily part of Juba politics rather than part of the local political processes in the state. Important decisions (like nominations of candidates) are, for example, made at the national level. The preferences of the government are thus more important than considerations of local contexts. Similar dynamics exist at the state level where the power of the governor is extensive (Ferrie Reference Ferrie2011). Hence, despite support in the constitution, decentralisation is generally absent and rules regularly do not take local conditions into account. Furthermore, within the SPLM the prospects for ordinary members to influence decisions are strictly limited, as party structures at grassroots level are lacking and decision making follows a clear top-down approach (Rolandsen Reference Rolandsen2007).

A final institutional problem is that boundaries of electoral entities are unclear and the introduction of elections in South Sudan has contributed to disarray over boundaries. South Sudan's Local Government Act stipulates that 70,000 inhabitants are needed for an area in order to constitute a county. If a region is smaller it is merged with a neighbouring county. A county entails a seat in the South Sudan Legislative Assembly, whereas a constituency size of 146,000 is needed for a seat in the National Assembly. This is stipulated in the Local Government Act from 2009, which is authorised by the SPLM (ICG 2009). This means that a redrawing of boundaries can be disastrous for a politician jockeying for power, because it can jeopardise his chances to safeguard a political position. Importantly, administrative changes can put a community that was previously in the majority in a constituency into a position of being a minority. With voting strongly following tribal lines in South Sudan, such changes are also likely to alter who will win an election in the area. Accordingly, administrative alterations are a useful tool in the fierce competition of the political sphere (UNMIS representative 2011 Int.). The communal conflict between Balanda and Dinka in 2012 illustrates these dynamics, as it was triggered by a decision by the regime in Juba to relocate the administrative headquarters of Wau County in Western Bahr el-Ghazal from Wau to Bagari. This verdict empowered the Dinka at the expense of the Balanda and 28 people were killed in the fighting (Sudan Tribune 27.12.2012).

To sum up, the institutional structures of South Sudan are too weak to handle political power struggles in a peaceful manner. Without formal institutions to manage conflicts through constitutional means – in particular in relation to the transfer of power – electoral politics continue to be waged with unconstitutional tactics, including violence.