Conflict Over the Commons: Government Bias and Communal Conflicts in Darfur and Eastern Sudan

Abstract
 Why do communal conflicts turn violent in some regions but not in others? This article identifies conditions for intercommunal cooperation and examines what makes such cooperation break down. Inspired by Ostrom's CPR-theories, it highlights three mechanisms—sanctions, boundaries, and local rules—underpinning intercommunal cooperation. Next, the argument stipulates that government bias can undermine conditions for communal cohabitation, tipping the balance in favor of conflict rather than cooperation. A systematic comparative study between Darfur and Eastern Sudan—building on extensive fieldwork—provides empirical evidence for the argument. These findings provide new insights on how to enhance community resilience to communal violence.


Introduction
In February 2015, an incident of cattle theft in South Darfur led to fierce fighting between two pastoralist communities-the Habaniya and Rizeigat Baggara-and the clashes killed several hundred people. When a Bergo farmer killed a man from the Fulani community in Eastern Sudan in 2009, people worried that this would spur fierce intercommunal fighting. 1 This did not happen. Instead, traditional local authorities quickly intervened, Diya (blood money) was paid and the situation soon calmed down. These examples are typical for these regions. In Darfur, communal disputes have often escalated to intensive fighting. Similar disputes with great potential for escalation have also been frequent in Eastern Sudan. However, they have generally been resolved before turning particularly violent.
The dynamics in one of the severest communal conflicts in Eastern Sudan are informative. In November 2009, a Hausa-Masalit conflict occurred in Gedarif. It started when Hausa youth beat up a Masalit minibus driver because he drove a young Hausa woman to her job as a domestic worker. The reason why this enraged the Hausa was a political editorial published in a newspaper claimed that the Hausa 'were not Sudanese.' If we are not Sudanese, they argued, our women should not work at Sudanese homes and no one should drive them there. The two communities brought people from the surrounding areas to the city and the tensions grew. When the government decided to arrest the Hausa's Omda (a high-ranking traditional leader), things got worse. Local leaders saw the arrest as a serious mistake as they feared that the shame of having the Omda arrested would escalate the conflict. However, the traditional authorities succeeded to convince the government to release the Omda and tensions eased. This was critical. A local academic asserted 'if the government had not released the Omda, there would not have been any more Gedarif'.
During the last three decades, communal conflicts, here defined as conflicts between non-state groups organized along a shared communal identity, 2 have killed thousands in Darfur but only a few dozen in Eastern Sudan. Why? This constitutes an empirical puzzle that existing theories cannot satisfactory explain. The two regions share several important features and scholars have described parts of Eastern Sudan as 'perfect mirror images of Darfur' (Babiker et al., 2005, p. 45). Eastern Sudan constitutes a theoretical irregularity: despite the prevalence of factors such as ecological degradation, political exclusion, and lack of development, intercommunal relations have remained largely cooperative. Both regions have also experienced civil war. Yet, whereas intercommunal relations remained relatively tranquil before, during and after the civil war in Eastern Sudan, violent communal conflicts were common in Darfur both prior to and during the war. The extensive sub-national variation observed in Sudan is typical for communal conflicts. Similar patterns exist in other countries heavily afflicted by this type of conflict such as Kenya, Nigeria, and Uganda (Sundberg & Melander, 2013). Whereas intercommunal cooperation distinguishes some regions, others are shattered by violent communal conflict that kill dozens, hundreds, or even thousands of people. 3 This article speaks to this variation. It does so by asking the following research question: Why do communal conflicts turn violent in some regions but not in others?
Communal conflicts pose a severe threat to human security and kill thousands of people each year. Nevertheless, it is more common that communities cohabit peacefully than fight with violence (Fearon & Laitin, 1996). This has theoretical ramifications. A theory of intercommunal violence needs to account both for conditions for cooperation and why such cooperation sometimes break down. This article develops such a theory. It first identifies conditions essential for underpinning intercommunal cooperation and then examines what makes such cooperation break down. Inspired by Elinor Ostrom's theories on common-pool resources (CPRs), this article highlights three mechanisms-sanctions, boundaries, and local rules-as essential for sustaining intercommunal cooperation. Next, the argument emphasizes how the state's conduct influences such cooperation. Governments regularly adopt different approaches to regions depending on the distinctive threats and opportunities a region presents (Boone, 2003). Such differences matter. In fact, when the government is biased, favouring some communities and disfavouring others, it is likely to undermine peaceful cohabitation. More precisely, in regions where the government pursues a biased agenda, such partiality will undermine cooperation mechanisms: making sanctions less efficient, land and administrative boundaries less clear, and rules less anchored in local conditions. Dynamics that are expected to exacerbate the risk of communal conflicts turning violent. A systematic comparison between Darfur and Eastern Sudan, drawing on interview material collected during six months of fieldwork in Sudan, supports these propositions. This article's prime contribution is developing and testing a detailed theoretical argument to explain why communal conflicts turn violent in some regions but not in others, and specifying how governments' behaviour influence intercommunal relations. That the state's conduct matters for how communities interact are not new. A government may, for instance, incite conflicts by providing weapons or by selectively directing the security forces. Yet, by pinpointing conditions crucial for intercommunal cooperation and then illustrating how the state can undermine these mechanisms, this study identifies new causal pathways concerning how the state influences intercommunal relations. Combining agency (the state's conduct) and an institutional perspective (mechanisms underpinning cooperation) results in a dynamic argument that reveals important insights about centreperiphery interactions.
The article also demonstrates that Ostrom's CPR theories, which originally concern non-violent relations between individuals or families, travel well to examinations of collective violence. While some theoretical studies reveal the potential of Ostrom's theories for probing collective violence (c.f. Ratner et al., 2013), empirical examinations of collective violence using this framework remain rare (for an exception see Oyerinde, 2019). An empirical asset is assessing Eastern Sudan, a region that previously has received limited attention. The extensive field research in remote areas of Sudan means that the article considers information from people who are often voiceless in academia. The article's comparative approach stands in contrast to most studies on Sudan that usually focus on one region. de Waal's (2016) examination of mass atrocities constitutes an important exception, but this article examines the violence of a lower magnitude and specifies mechanisms critical for cooperation and conflict at the local level. The findings presented in the article provide new insights on how to enhance community resilience to communal violence.

Previous Research and Motivation
Our understanding of communal conflict has improved significantly during the last decade. One important strand is quantitative studies assessing communal conflict across countries (c.f. Fjelde & von Uexkull, 2012) or within a single country through a grid cell approach (c.f. Theisen, 2012). Qualitative analyses of communal conflict in a single case (c.f. Krause, 2019) have also provided important insights. Studies using a systematic qualitative comparative approach are fewer. While there exist such studies examining intercommunal relations in relation to elections (c.f. , or in the wake of authoritarian regime breakdown (c.f. Krause, 2018;Tajima, 2013;Van Klinken, 2001), we know less about how intercommunal variations vary without such turbulent dynamics. This article contributes to fill this lacuna by examining communal conflicts under Omar al-Bashir's 30 years in power . While elections existed, they were relatively uncompetitive and none of the examined conflicts were closely related to electoral dynamics.
Examining a government's approach towards two large regions for a long time period obscures some nuances. However, it also enables an adequate exploration of how the government's conduct differs between regions and how this impacts intercommunal relations. It also allows for careful examination of the case where communal conflicts did not turn violent. Previous studies generally compare violent and non-violent communal conflicts without unpacking the latter (c.f. Fjelde & von Uexkull, 2012;Hillesund, 2019). Lumping together cases with harmonious relations and cases where disputes are widespread, but solved before escalating, has theoretical ramifications. Although conflict triggers (such as cattle raiding or land disputes) are important for the onset of conflicts, examining variations in violence following a dispute requires different explanations.
Darfur and Eastern Sudan share many features identified to increase the risk of communal conflict. One research strand concerns climate factors and emphasizes that resource scarcity, droughts, soil degradation, and environmentally-induced migration increase the risk for such conflicts (Fjelde & von Uexkull, 2012;Homer-Dixon, 1999). Yet, Darfur and Eastern Sudan face similar ecological degradation (Elhadary & Samat, 2011;Taha, 2007), including persistent drought, land degradation and shrinking pasture areas (UNDP, 2010). In rural areas, communal conflicts generally occur between pastoralists and farmers, or among pastoralists. While the percentage of pastoralists is higher in Eastern Sudan than in Darfur (DRDC, 2010), farming, and herding (of cattle and camels) are central livelihoods in both regions and livelihood disputes frequently occur in both areas.
A related argument emphasizes that economic and political exclusion increase the risk for intercommunal violence (Fjelde & Østby, 2014;Hillesund, 2019). The regions examined here are relatively similar in terms of political exclusion and lack of development (Young, 2006). The Black Book, 5 for example, reveals that both Eastern and Western Sudan are severely marginalized on indicators such as literacy rates and infant mortality (Cobham, 2005). According to some figures, Eastern Sudan has Sudan's highest poverty rate (AFP, 2012). The presence of refugees could increase resource competition (Böhmelt et al., 2019;Fisk, 2019). Yet, both Darfur and Eastern Sudan host large refugee populations and several refugee camps (UNHCR, 2019). Ethnicity is another central factor for many communal conflicts. However, the regions' ethnic diversity is comparable and most ethnic groups inhabiting Darfur are present in Eastern Sudan (Al-Hardallu & El Tayeb, 2005b). Patrimonialism characterizes most countries experiencing communal violence (Bates, 2008;Lemerchand, 1972). In this context, manipulative political elites can intensify conflict fought over scarce resources (Greiner, 2013;Turner, 2004). However, our knowledge about why such destructive manipulation takes place in some areas but not in others remains limited.
Another cluster highlights institutions. Local institutions are fundamental for interethnic cooperation (Fearon & Laitin, 1996) and for determining whether natural resources turn into a blessing or a curse (Adano et al., 2012). While revealing the significance of institutions, important components remain unaccounted for. Fearon and Laitin emphasize equilibria important for understanding interethnic cooperation, but do not address where these equilibria are most likely to occur, or the conditions promoting the equilibria. A vibrant intercommunal civic life decreases the risk for ethnic riots (Varshney, 2002). The opportunities, however, for such an animated intercommunal civic life differ from an urban context (Varshney's focus) to a rural setting; pastoralists' nomadic lifestyle impedes such daily intercommunal interaction. Examining ethnic conflict over land, Boone (2017) illustrates the importance of land tenure institutions. Areas with neocustomary systems are less prone to violent conflict compared to those with statist land tenure. Yet, land tenure is overall neocustomary in both Darfur and Eastern Sudan. Competing customary and modern jurisdiction can increase the risk of communal violence (Eck, 2014) but cannot explain subnational variation under the same jurisdiction. Furthermore, outside actors can facilitate intergroup dialogue (Smidt, 2020) but such initiatives are common in Darfur, not in Eastern Sudan.
Other studies examine how the strength and nature of local institutions shape local conflict dynamics. Wig and Tollefsen (2016) find that local institutions of high quality have a pacifying effect in a civil war context and Mustasilta (2020) reveals that contestation over traditional authorities reduce the efficiency of local institutions and thereby increases the risk of protest. This article agrees that the quality of local institutions matters. It complements existing studies by explicating how the state's conduct influences local institutions and by presenting a causal story that explains why the quality of local institutions may differ between regions and over time.
Traditional authorities constitute a viable complement to state-centred approaches (Mac Ginty, 2008) and are an important arena for resolving conflicts in large parts of Africa (Logan, 2013). Incorporating customary authorities into the state structure lower the risk of armed conflict (Mustasilta, 2019). Furthermore, groups with more formalized customary institutions are less likely to partake in communal conflicts (Wig & Kromrey, 2018) and civil war (Wig, 2016). Traditional institutions are essential for inter-communal mediation across Sudan (Leonardi & Abdul-Jalil, 2011). Formalized customary institutions exist in both Darfur and Eastern Sudan. Interestingly, Darfur's largest ethnic group, the Fur, is the Sudanese group with the highest score of formalization in Wig and Kromrey's (2018) data.
Despite the similarities between Darfur and Eastern Sudan, significant differences exist. The regions have, for example, experienced civil war to different extents and during different times. While civil war has shattered Darfur since 2003, Eastern Sudan only experienced a limited insurgency in this same period. When Eastern Sudan experienced civil war 1996-2001, no such conflict occurred in Darfur. This study considers how civil war has affected communal conflict. Furthermore, modern weapons are more abundant in Darfur (partly due to spillover from Libya and Chad). However, some communities in Eastern Sudan have access to modern weaponry and traditional weapons (which can cause extensive death tolls) are widespread in the region.
In sum, existing theoretical notions do not explain the empirical puzzle at hand. We hence need a better understanding of why communal conflicts become violent in some regions but not in others and how the state and institutions influence intercommunal relations. Next, I develop a theoretical argument for approaching these questions. It first outlines conditions underpinning communal cohabitation and then elaborates how government bias may undermine such cooperation.

Conditions Underpinning Cooperation
The CPR theory delineates conditions enabling cooperative management of shared resources, such as land, and elucidates how local institutions may enhance cooperation. Factors conducive to cooperation include trust, communication, and accurate information. Previous dealings are vital in building trust: the chances of reciprocity increase when a custom of mutual reliance and shared concerns exist. In contrast, levels of cooperation are lowest when the other is unknown, anonymity guaranteed, and opportunity to build reputation non-existing (Ostrom, 2008;Poteete et al., 2010). What are the more precise conditions promoting cooperation in a CPR setting? After analyzing the wide variety of successes and failures of management of CPRs, Ostrom exposed some regularity and came up with eight design principles 6 that enhanced management of the CPRs. Fundamental to these design principles is that they promote cooperation under difficult circumstances (Ostrom, 2008). It is this aspect of the design principles, rather than the exact organization of the CPR, that are important for this article. Not all design principles carry equal importance in a communal conflict context. For analytical purposes, the design principles considered most relevant for communal conflict have been modified to the topic in this article and structured into three mechanisms capturing aspects important for inter-communal cooperation. 7 All three mechanisms are assumed to independently influence intercommunal relations. Interactions between the mechanisms may exist, but in this article, I refrain from stipulating any theoretical predictions on this point.
The first mechanism-sanctions-contests that sanctions against perpetrators of communal violence can be an efficient tool for solving intercommunal disputes. However, for sanctions to enhance cooperation and boost trust they have to reflect the severity and context of a violation because communities are more likely to perceive such sanctions as fair, which, in turn, increase the chances that they will adhere to them. Conversely, if other factors, such as a perpetrator's communal affiliation, outweigh this notion of proportionality, certain communities will receive selective impunity. This will impair trust and render intercommunal cooperation less likely. In addition, selective impunity decreases the disincentive to resolve disputes with violence among groups not having to fear punishment. This reduces the chances of intercommunal cooperation and increase the risk of revenge that might lead to vicious circles of violence.
The second mechanism-boundaries-centres on who has the right to access a particular resource. Disputes over land-use rights are commonplace among pastoralists and between agriculturalists and pastoralists. Clearly demarcated boundaries provide an undisputed reference point that make solutions around land-use conflicts more likely. Longstanding systems ascribing a certain area to a particular community, and stipulating pastoralists' movements with their herds, have traditionally regulated land disputes. Changes to boundaries often disregard traditional perceptions and can invoke discrepancies between current and historic boundaries that communities tend to perceive as the 'real'. This study considers both administrative structures and land-use borders because these are often closely related. Administrative boundaries can cut off natural resources from a community that previously had access (Adano et al., 2012). Disputes over local administrative units are particularly prone to become salient if adjustments change the balance of power between groups (Posner, 2004). A community disfavoured by such alterations might take up arms to make up for its subsiding influence.
The third mechanism-local rules-emphasizes that it is critical for cooperation that rules reflect local conditions and that communities affected by the rules have the ability to modify them. Communal conflicts often take place in distant areas where communities habitually have developed rules to manage intercommunal coexistence. This system builds on traditional knowledge with detailed conventions on how to uphold peaceful relations. Rules anchored in local conditions increase familiarity with them and enhance cooperation. Yet, if communities feel alienated from the rules, the incentive to follow them declines. Proper understanding of local dynamics thus enhances cooperation. Theoretically, outside actors can enhance or obstruct the chances of cooperation among local communities (Ostrom, 2008), but in practice, they commonly fail to understand social customs and norms among groups. It hence conduces cooperation if procedures of monitoring are internal, rather than external. Furthermore, monitors are more efficient if they are accountable to community members as this increases people's trust in them (Dolšak & Ostrom, 2003;Ostrom, 2008). Local situations often change rapidly; rules are more efficient if affected people can make apt modifications when conditions change. Additionally, local community leaders generally know how best to communicate rules to maximize cooperation.

Government Bias Undermining Cooperation
External actors can influence the conditions underpinning cooperation. The government's conduct is central because it has the ability to influence several factors important for intercommunal relations. The government's strategic interests are significant when deciding whether it should intervene or not in a communal conflict (Elfversson, 2015; and biased decisions concerning property-rights increase the risk for pastoralist conflicts (Butler & Gates, 2012). In addition, actors at the centre might ally with actors involved in local conflicts, which can generate conflict by aggravating elite interactions between central and local elites as well as among local elites (Kalyvas, 2003).
This article argues that government bias-the extent to which a regime differentiates between communities by favouring some and disfavouring others-are fundamental for conditions underpinning cooperation. Government bias is not a dichotomous variable and may shift between different parts of the regime. This article concerns the general level of partiality. In a context where the central authority acts with relative neutrality, opportunities for peaceful cohabitation remain. In contrast, when partiality permeates the government's conduct it is likely to destabilize conditions underpinning cooperation.
First, a biased government can undermine the effectiveness of sanctions. Such partiality could make sanctions depend on the communal affiliation of the perpetrator (or victim) rather than reflecting the gravity and context of a violation. A partial regime may also give impunity to certain communities and their leaders. This has a dual effect. It reduces the disincentive for favoured groups to use violence (as they can count on immunity) and creates frustration among disfavoured communities, which may push them to settle disputes outside the legal framework as jurisdictive decisions are unlikely to be fair. Second, government bias is likely to result in less clear boundaries. A partial government may change boundaries to weaken communities it perceives as threats, or to compensate communities that have supported it. Such alterations will create disarray: not only does it increase uncertainty in determining who controls a particular area, but it can also shift power balances between communities. Third, a biased government is likely to make rules less contingent on local conditions and restricts the influence of local actors affected by the rules. Rather than ensuring that rules reflect local circumstances, biased policies from the government will seek to ensure that rules reflects their interest. To minimize the risk of decisions not following its agenda, a biased regime may also restrict local actors from certain communities to participate in modifying rules. A partial agenda is also likely to promote local leaders (both in traditional and governmental structures) that prioritize the regime over the local communities' interests, which further distance rules from local conditions. Thus, government bias can undermine conditions underpinning intercommunal cooperation. As such, we can expect violent communal conflicts to be more prevalent in a region rife with biased state interaction than in a region spared from such dynamics.

Methodological Approach
To evaluate the explanatory power of the causal story, this article combines within-and between-case analysis, a suitable approach for evaluating the inferential value of the independent variable and for assessing alternative explanations (Bennet & Checkel, 2015). In line with the advice in case-study methodology (c.f Beach & Pedersen, 2019;George & Bennet, 2005), I have strategically chosen regions where the outcome varies and compare most similar cases, as it is suitable for identifying mechanisms omitted by previous research. I use a structured and focused comparison method to compare the regions. It is structured in that the questions asked are the same to all cases and it is focused in that it only deals with certain aspects of the cases (George & Bennet, 2005).
A prerequsite for a valid comparison is a proper independent analysis of the cases. In fact, comparative case studies should not be pursued without extensive case knowledge. Deep comprehension of the cases is essential to avoid an overly mechanical application of a comparative analysis and important to detect previously overlooked factors. Within-case analysis are essential for causal inference (Beach & Pedersen, 2019;Mahoney, 2010). In this process, it is important to look for causal-process observations (CPOs).
A causal-process observation is an insight or piece of data that provides information about context or mechanism and contributes a different kind of leverage in causal inference … a causal-process observation may be like a ''smoking gun.'' It gives insight into causal mechanisms, insight that is essential to causal assessment and is an indispensable alternative or supplement to correlation-based causal inference. (Collier et al., 2010, pp. 184-185) A pertinent way to discover CPOs is through intensive data gathering of key events, for instance through field research (Mahoney, 2010). 8

Field Research
Interviews carried out during six months of field research in 2007-2015 constitutes the empirical foundation for the article. Some interviews were group discussions; I talked to 133 persons during the 114 interviews. Of primary interest were stakeholders with firsthand information about intercommunal relations. This included farmers and pastoralists, representatives of pastoral and agricultural unions, traditional authorities, and government officials. To contextualize and verify this information, I gathered data from academics, NGO and IGO representatives, and other secondary sources. I complemented interviews in Darfur (Nyala) and Eastern Sudan (Gedarif and surrounding areas) with interviews in Khartoum and Juba (South Sudan). The security situation in Darfur meant that it was easier to talk to Darfurians outside the region.
Field research in Eastern Sudan facilitated a thorough analysis of the understudied 'noncase', evaluation of the theoretical framework and probing the dependent variable. While information about disputes that turned violent often exist, we know little about the cases that did not. Field research in Eastern Sudan provided a comprehension of why disputes did not escalate in the region. Secondary sources revealed only one violent communal conflict in Eastern Sudan. To decrease the risk of overlooking any conflict, I consulted customary authorities and local academics during a seminar at the University of Gedarif. The participants were, after a lively discussion, in agreement that no other conflict in Eastern Sudan met the article's criteria for a violent communal conflict.

Darfur: Government Bias Disrupts Cohabitation
Darfur is a marginalized region in western Sudan. In 2003, two Darfurian insurgent groups (Sudan Liberation Movement/Army, SLM/A, and Justice and Equality Movement, JEM) launched a rebellion against the Sudanese government. The origin of the two groups differed. The SLM/A's roots are primarily within self-defense forces within three non-Arab communities (the Fur, Zaghawa, and Masalit) created for protection against assaults by Arab militia groups. JEM, in contrast, originated from the centre of power and was to a large extent the result of a split in the Islamic Movement (Roessler, 2016). The government's response included the use of militias, induced famine, and scorched earth tactics (Flint & de Waal, 2008). The conflict has resulted in hundreds of thousands of fatalities (from direct violence, malnutrition, and diseases) and displaced millions (de Waal, 2016). The US administration (and others) considered that the atrocities committed by the government and affiliated militias constituted genocide (Nathan, 2007).
The 2003-rebellion was not the start of conflict in Darfur. The region has long suffered from conflicts stemming from a combination of: local conflicts, the government's extensive use of militias, and incursion from rebels based in Southern Sudan. These  . 9 It records 20 violent communal conflicts in Darfur in 1989-2018, estimated to have killed between 7300 and 8500. Table 1 charts ethnicity, livelihood and fatality estimates for these conflicts. Yet, ethnicity is fluid, livelihoods often mixed, and fatality data difficult to confirm. Many label the Darfur conflict as an African-Arab conflict. This dichotomy is far too simplistic and lacks nuance; the base for this division is not language, skin colour, or religion (all Darfur's communities are Muslim). Rather, this separation is founded on group claims to Arab identity. Identity, however, is a powerful force-and for those who claim it, the Arab identification is vital (Tubiana, 2007). Two-thirds of the conflicts stood between pastoralists groups and one-third pitted farmers against herders. Groups that both identify as Arabs fought 11 of the 20 and 9 conflicts stood between Arabs and Africans. The pattern of who is fighting who has changed since 2003. Before the rebellion, most conflicts were between Arabs and Africans. After the rebellion, a majority has pitted Arabs against Arabs. This shift relates to land. The government's counterinsurgency forced millions of non-Arabs to leave their land and inter-Arab conflicts over this land followed (Brosché & Rothbart, 2013).

Countering Threats and Unleashing Militias
A strong partiality has permeated the Sudanese government conduct toward Darfur's communities throughout the examined period. The regime has strongly disfavoured three non-Arab communities (the Fur, Zaghawa, and Masalit) and favoured Arab groups. This policy is partly based on promotion of Arabization (Burr & Robert Collins, 2008). In 1990, Sudan signed a deal with Libya (to secure weapons supplied by Qaddafi) that promoted the Arab culture in Darfur (Lesch Mosely, 1998). The government's conduct was, however, primarily based on calculations concerning threats and opportunities. Khartoum perceived some communities as enemies and strongly relied on militias to fight against these groups. To facilitate recruitment into these militias, it has favoured certain communities. Militia tactics sometimes create chaos, but this has not restricted the government. Although Darfur offers some economic assets (such as livestock and, in recent years, gold), the region has not been economically vital for the government. The government's biased conduct is also entrenched in its ambition to keep local actors busy with local struggles. Khartoum engaged divide-and-conquer tactics to promote tribalization and aggravate tensions between, and within, Darfur's communities (Tubiana et al., 2012). 10 When el-Bashir took power in 1989, the war in southern Sudan was the government's largest threat, and the regime worried that it would spread northwards (Flint & de Waal, 2008). A year later, Daud Bolad, a Fur Islamist close to the government, defected to the Sudan People's Liberation Movement/Army due to the regime's partiality in the Fur-Arab war. Bolad launched an insurgency in Darfur. The government perceived this as a Fur insurgency and started to view the Fur community as its main enemy in Darfur. To counter the rebellion, the regime recruited militias from some Arab communities. This both cemented the government's alliance with these groups and increased its bias against 'anti-government' communities (ICG, 2004). The government and their militia allies, partly owing to established networks between regional and central Islamists (Roessler, 2016), quickly defeated Bolad's rebellion. In 2003, the insurgents primarily came from non-Arab communities (although it had some Arab inclusion). 11 The government again recruited Arab militias, which intensified the regime's favouritism of communities from which it could glean recruits (Flint & de Waal, 2008).
This pattern of government partiality toward Arab communities does not imply that all government decrees disfavoured non-Arabs. Sudan's Islamic movement (a key component of Bashir's regime) has included important Darfurian non-Arab elements since the 1960s. Initially, devotion to Islamism overshadowed ethnic belonging. After ascending to power, however, competition over state resources intensified ethnic divisions. When the Islamic movement split in the late-1990s, Bashir's faction strongly favoured Darfurian Arabs to counter the strong support that the rival faction (led by Hassan al-Turabi) had from many Darfurian non-Arab Islamists (Roessler, 2016).

Selective Sanctions Contribute to Violence
Fair sanctions underpin cooperation. In Darfur, however, accountability for crimes has been determined more by group membership than the severity of the violation-a situation that largely stems from the government's bias.
Historically, traditional institutions and customary courts have been crucial for fostering constructive intercommunal relations in Darfur. A key component has been determining sanctions in intercommunal disputes. The effectiveness of these institutions rested on trust in the system and neutrality over their hearings (Mohamed, 2009). The government traditionally had a facilitator role but refrained from direct involvement. Since 1989, however, the regime has taken an active role in customary institutions, which has reduced their influence and undermined their efficiency (Tubiana et al., 2012). 12 Those communities considered pro-Khartoum by the government generally received preferential treatment. A central component for resolving intercommunal conflicts in Darfur are the payment of Diya (blood money). 13 Diya payments are a form of restitution-often instituted as part of a communal reconciliation process-that help to prevent cycles of violence after serious interpersonal crime such as murder (Fallon, 2020). When conflicts became more acute in the late 1980s, Diya amounts increased, and the government promised to pay the balance if the involved parties could not meet the expense. Yet, not all communities received this equally; those seen as hostile did not receive Diya. In the mid-1990s, for example, the government did not provide Diya to the Zaghawa after a conflict between Abbala Arabs and the Zaghawa. As a consequence, peace initiatives failed and violence resumed (Tubiana et al., 2012).
The government has also undermined sanctions by selectively providing amnesties and impunity. This has contributed to revenge attacks and decreased disfavoured groups' incentives to seek legislative solutions. In relation to Daud Bolad's rebellion, the government ignored severe atrocities committed by Arab groups against the Fur, which frustrated the latter and destroyed the social fabric for intercommunal cooperation (Flint, 2007). The government's exemption from sanctioning certain communities also contributed to the Arab-Masalit conflict in the mid-1990s. The government's inaction against Arab raiders that destroyed villages and killed many people infuriated the Masalit community and escalated the conflict (Flint & de Waal, 2008).
The 2003-rebellion further diluted sanctions and resulted in a proliferation of amnesties (HRW, 2007). For example, the government released the infamous Janjaweed leader Musa Hilal from prison to empower its counterinsurgency (Roessler, 2016). Not only did this affect the civil war, but it also intensified intercommunal conflicts because Hilal used his weapons, position, and impunity to procure land for his landless community (HRW, 2007).

Unclear Boundaries Disrupt Cooperation
Land is economically, politically and socially essential in Darfur and disputes over land rights are frequent (Unruh & Abdul-Jalil, 2014). The Dar (homeland) system provides access to land and political recognition. A Dar defined the area, habitually inhabited by different ethnic groups, under which a paramount chief (most often called Nazir) had authority. This structure generally favoured larger communities, habitually in control of Dars, over smaller groups (Leonardi & Abdul-Jalil, 2011). Historically, this arrangement facilitated coexistence between Darfur's communities but the government's increased interference has disrupted the system (ICG, 2004). As part of its divideand-rule strategy, the regime has repeatedly altered boundaries to weaken 'anti-government' communities and to secure the support of others. A relatively structured land system with transparent rules and demarcations was replaced with a system that introduced uncertainties and risk, and incentivized opportunism (Tubiana et al., 2012). Uncertainty over land rights makes conflicts more frequent. It also increases the stakes involved and the risk for violence (Unruh & Abdul-Jalil, 2014). The split of Darfur into three states in 1994 constitutes an important example of the government's strategic use of boundaries. In fact, this was 'perhaps the most crucial decision' (Burr & Robert Collins, 2008, p. 287) for the disastrous developments that followed. The division sought to weaken certain communities, particularly the Fur, traditionally supporters of the Umma party, the ruling National Congress Party's (NCP) prime political rival in Darfur (Roessler, 2016). The division split the Fur's traditional stronghold, Jebel Mara, into all three new states-seeking to dilute the Fur's relative power in Darfur. The Fur went from a majority position to a minority position in each state (ICG, 2004). In 1995, another administrative adjustment by the government divided the Masalit's traditional homeland into thirteen domains, whereof five were allocated to Arab groups. This amendment infuriated the Masalit and triggered an Arab-Masalit conflict (ICG, 2004).
Boundaries have also been important for several land-related Arab-Arab conflicts. To reward the communities that provided the strongest support during the counterinsurgency, the regime selectively provided land. This created uncertainty over boundaries and aggravated inter-Arab enmity (ICG, 2007). The Maaliya-Reizegat Baggara conflict centres on the Maaliya's desire for a Dar, which the Reizegat Baggara has denied since it would be carved out from their land (Mamdani, 2009). However, to weaken the Reizegat Baggara (who were not a strong government supporter) the government favoured the Maaliya and gave them a Dar in 2003, which escalated the conflict between the two communities (Tubiana et al., 2012).

Disregarding Local Rules Destroys Conflict Management
Many Darfurians have stressed the importance of considering local rules. 14 Yet, the government has actively weakened local structures and circumvented local actors. A statement by the deputy governor of South Darfur is illustrative: We want to start from scratch, to stop using traditional mediation. No judiya [traditional justice and reconciliation, explanation added] or traditional court. Law only. We want to impose the law, otherwise people will not feel that there is a government. (cited in Tubiana et al., 2012, p. 83) As part of this policy, the government proliferated peace conferences. Yet, these government-sponsored initiatives have generally failed to promote peace because local conditions were unconsidered. 15 The regime has also restricted influential local actors by disregarding customary power in appointments to traditional authorities. Thus, the appointees do not genuinely represent the communities; their prime allegiance is to the regime. 16 Not adhering to local rules had grave consequences for the Arab-Fur conflict. In May 1989, the communities should have met in a conference: a well-established tradition of consultation recognized by all ethnic groups. However, emboldened by strong backing from the government, the Arab communities decided 'in an astonishing disregard of these established customs' (Burr & Robert Collins, 2008, p. 244) to boycott the conference. Fighting resumed within days. Lack of consideration for local circumstances also influenced the Arab-Masalit conflict. To empower the Arabs, the government entitled newly appointed Arab administrators Amir, a designation that customarily only the son of the Sultan (paramount Masalit chief) was entitled to. This enraged the Masalit and aggravated Masalit-Arab relations (Osman, 2006).

Eastern Sudan: Impartiality Reinforcing Cooperation
Eastern Sudan, a frontier area bordering northern Ethiopia and western Eritrea, has witnessed conflicts for decades. The regimes in Khartoum, Addis Ababa, and Asmara have repeatedly supported one another's opponents (Young, 2006). The largest ethnic group in Eastern Sudan is the Beja, a non-Arab community divided into various sub-groups. In numbers, two groups identifying as Arabs (the Shukriya and Dabyna) follow. Many other Arab and non-Arab groups, as well as numerous communities originating from South and West Sudan, also inhabit Eastern Sudan (Al-Hardallu & El Tayeb, 2005a). As part of the anti-government umbrella organization National Democratic Alliance (NDA), the Beja Congress initiated an insurgency in 1995 (ICG, 2006). The NDA-rebellion continued until 2001, and the fighting caused more than 3000 deaths (UCDP, 2021). In 2003, the Beja Congress restarted its military action, and together with the Rashaida Free Lions, formed the Eastern Front in 2005. After a few years of low-intensity rebellion, fighting ceased in 2006 with the signing of the Eastern Sudan Peace Agreement (ICG, 2013).
In rural Eastern Sudan, small-scale agriculture, pastoralism, and mechanized farming are the three dominant livelihoods. Disputes between pastoralists, and between pastoralists and farmers, over land and water are common. 17 Sometimes groups challenge other communities to demonstrate their political importance. 18 Strong collectivist identities in Eastern Sudan sometimes result in individual conflicts turning communal (Al-Hardallu & El Tayeb, 2005b). The conflicts have not been particularly violent (el-Amin 2004). The disputes have occasionally caused a few fatalities with about 10−15 persons dying yearly. 19 Thus, few disputes have met this article's threshold-25 fatalities in a calendar year-for a violent communal conflict. In fact, the only exception during al-Bashir's 30-year rule stemmed from groups who had fled to Eastern Sudan from Southern Sudan. In 1997, a Nuer-Dinka conflict killed 35 people; this violence was rooted in conflict dynamics from outside the region of interest (UNHCR, 2010).

Protecting Resources and Avoiding Chaos
Throughout the studied period, Khartoum's policy towards Eastern Sudan has not been entrenched in partiality towards certain communities. The apparent differentiation between Arabs and non-Arabs in Darfur was not a distinctive factor for the government's policy in Eastern Sudan. People interviewed in Eastern Sudan overwhelmingly expressed that the government did not differentiate strongly between communities: no community was consistently and strongly disfavoured or favoured. Yet, attitudes toward the government were not positive. Instead, interviewees from various communities repeatedly noted Eastern Sudan's staggering poverty-which cuts across ethnic affiliation-as the prime complaint. Despite the general perception of relative impartiality, many pastoralists held that the government somewhat favoured the agriculturalists over them (Babiker et al., 2005). 20 Throughout al-Bashir's rule, Eastern Sudan was economically and strategically vital. Oil constituted the preeminent economic asset, and Port Sudan was Sudan's only oil-exporting harbour (ICG, 2006). Illustratively, an insurgency attack in 1997 aimed to 'cut off Port Sudan, which is the lifeline of Khartoum' (Reuters, 1997). To secure control over Eastern Sudan, Khartoum mobilized an extraordinary display of military strength and has maintained a high-security presence (ICG, 2006). 21 This strategy sought to promote stability. A chaotic Eastern Sudan, resembling the situation in Darfur and Southern Sudan, would be disastrous for the government. Ethnic militias, they believed, could jeopardize stability. In sharp contrast to other areas of Sudan, the government has been reluctant to employ such tactics in Eastern Sudan. 22 In the few attempts to recruit militias in Eastern Sudan, the regime has primarily sought people to fight outside the region (ICG, 2013) and has generally not provided high powered weaponry or significant influence (ICG, 2006). 23 The regime's preferred tactic to promote stability has been co-optation (ICG, 2006). As part of this strategy, the government did not strongly differentiate between communities but sought to establish relations with leaders across groups. Importantly, the government deemed that prominent communal leaders could contribute to the desired stability, and has not sought to undermine or replace such actors (Assal, 2013). This created a mutual dependence where the government used the influence of local leaders to keep stability and these communal elites depended on the government for their influential positions. 24 A religious leader in Gedarif provided a telling summary.
The government cannot solve many problems and instead does it through contacting the Native Administration. In general, the government is neutral and wants to solve the problem but sometimes a government official might not be neutral but support his tribe. 25

Fair Sanctions Counters Escalation
In Eastern Sudan, sanctions have ultimately depended on the seriousness and context of a violation, rather than factors such as communal affiliation. The government's relatively neutral conduct meant that no community had blanket impunity and no community was singled out as an enemy constantly ruled against. 26 Traditional local leaders have historically been important for settling communal disputes in Eastern Sudan. In contrast to Darfur, the government has not undermined customary institutions in Eastern Sudan and traditional authorities have remained influential. A committee for conflict resolution, comprised of elites from various communities, meets when a dispute emerges. When the Beni Amer-Beja conflict started in Port Sudan, local traditional leaders from the whole of Eastern Sudan gathered and prevented escalation. 27 The government and customary leaders often cooperated to manage conflicts. Sometimes, the regime imprisoned criminals identified by the customary authorities. 28 A traditional institution with wide representation, together with a fairly unbiased government, ensured that sanctions were relatively neutral. Sanctions for a violation were determined by the severity of the violation itself, not group membership of the victim or the perpetrator. This increased communities' confidence that rulings were fair and promoted cooperation. 29 People interviewed in Eastern Sudan consistently hailed local traditional leaders as fundamental for the largely peaceful inter-communal coexistence, and attributed such leadership to the non-escalation of conflict. In fact, even people who generally opposed the customary system (because they saw it as an outdated system that excluded young people and women) still considered it crucial for land issues and intercommunal disputes. A prerequisite for the effectiveness of this institution was that the government did not infringe.
Trustworthy and neutral mediators are key to the successful outcome of conflict resolution processes [in Eastern Sudan, explanation added]. Such mediators are only effective if they are acknowledged as neutral parties and not if they appear to be representing a stakeholder in the conflict (including a particular ethnic group or government interest) [italics added]. (Al-Hardallu & El Tayeb, 2005b, p. 6) Appropriate sanctioning has contributed to the settlement of numerous disputes in Eastern Sudan. In November 2008, a Beni Amer pastoralist killed a Masalit farmer in Gedarif, and people worried that this would spur an intercommunal war. However, community leaders from many different ethnic groups as well as the Wali (governor) went to ease the situation. The government imprisoned the perpetrator, which convinced the Masalit not to retaliate. The conflict did not escalate, and no additional killings occurred. 30 The Diya system, an important device for peaceful coexistence throughout Sudan, depends on fairness. If communities perceive a Diya decisions as prejudicial, it loses its credibility. In Eastern Sudan, such decisions are generally perceived to be fair across communities. As such, this tradition has remained important to settle communal disputes, and to prevent small scale violence from escalating. For instance, after a Bergo farmer killed a Fulani man in 2009, local traditional authorities instantly arbitrated, Diya was paid, and the situation resolved. 31

Undisrupted Boundaries Underpin Collaboration
Land is essential for livelihood in Eastern Sudan. Ownership of land symbolizes cultural identity and is central for political recognition. As in Darfur, land is organized in different Dars and competition for Dars has triggered disputes. These have, however, not escalated to violent communal conflicts (el-Amin 2004;UCDP, 2021). A group that desires a Dar is the Rashaida but the Beja, who are not only economically dependent on their land, but also culturally and emotionally attached to it, have thwarted such attempts (Al-Hardallu & El Tayeb, 2005b). In 1994, the government gave the Rashaida a Nazirate. Although a Nazirate did not provide land ownership, it amplified the political recognition of Rashaida. The verdict thus disrupted power relations less than granting a Dar would have. While this decision left the Beja dissatisfied (ICG, 2006), it has not led to an outbreak of violence (el-Amin 2004;UCDP, 2021).
Administrative boundaries in Eastern Sudan have remained fairly clear and administrative adjustments have not (in sharp contrast to Darfur) sought to alter power balances or to aggravate local tensions. 32 For example, the split of Eastern Sudan into three states in 1994 did not significantly shift power balances between the communities because it was not designed to favour some and disfavour others. 33 Administrative units have thereafter remained rather stable. 34 Unclear demarcation of masarats (roads used by pastoralists when moving their animals) has however caused farmer-herder disputes. The government has allocated these roads to the pastoralists. However, they have neither communicated this to the farmers nor purchased the land from them. Both groups view themselves as rightful owners of the land. 35 While the pastoralists complained that the masarats were too narrow and that farmers often cultivated them, the farmers' key dissatisfaction was pastoralists' animals eating their crops. 36 These disputes are important for livelihood, but have not prompted any violent communal conflict.

Local Understanding Enhances Conflict Management
In Eastern Sudan, the government has generally tolerated rules relating to local circumstances and permitted local actors to modify regulations. Islam is deeply rooted in the region and rules with a foundation in Islam are less likely to be broken than those disregarding religion (Al-Hardallu & El Tayeb, 2005b). The devotion to Islam has been important to manage intercommunal relations. When a dispute emerged, meetings to settle it were often held in mosque and verses of the Quran quoted to emphasize how Muslims should deal with conflicts. 37 Rules were further locally anchored by prominent local elites with a refined understanding of the local context. Instead of undermining such actors, the government allowed them to be influential. The government had, in fact, committed itself to consult with customary authorities if an intercommunal killing took place in Eastern Sudan. 38 Traditional authorities also observed how the government applied rules (for instance regarding trespassing) to ensure that local conditions were considered. 39 In Kassala state, a council where traditional local leaders gather to solve conflicts has been successful. While the government financed the project and maintained a symbolic presence, the customary authorities controlled the work (Al-Hardallu & El Tayeb, 2005b). This arrangement contributed to cooperation as the communities were more likely to accept regulations influenced by traditional leaders than if they came solely from the government. 40 Furthermore, the government sometimes adjusted its decisions after consulting customary institutions. An example from Gedarif illustrates how imperative such modifications can be. In November 2009, a Hausa-Masalit conflict caused 13 fatalities-making it the severest communal conflict in Eastern Sudan in many years. 41 When state authorities arrested the Omda of the Hausa, many Hausa (and local academics) interpreted this as evidence of the government's partiality. Traditional leaders feared that the arrest would escalate the conflict, and succeeded to convince the government to release him. 42 Hence, the government altered its decision despite ostensible favouritism toward the Masalit, which deescalated the situation. 43

Conclusions
This article set out by developing a theory about communal conflicts and how the government's conduct influences intercommunal relations. The Darfur-Eastern Sudan analysis provided empirical validation to the theoretical argument. The policy towards the two regions differed starkly under al-Bashir's rule. Partiality permeated the regime's conduct to Darfur and it strongly favoured 'pro-government' communities and disfavoured 'anti-government' groups. In contrast, the government acted with relative impartially towards communities in Eastern Sudan. The difference rested on a risks and opportunities logic. Because the government was not dependent on resources from Darfur, its response to the threats it encountered from that region could rest on strategies that often evoke chaos, such as the use of militias and divide-and-conquer tactics. This was not a good alternative for Eastern Sudan, a region of the highest economic and strategic importance for the regime. An important difference was that while the government strongly favoured Arabs over non-Arabs in Darfur, such a distinction was not apparent in the east. This finding nuances the discussion about how central 'Arabization' was for Bashir's government. It certainly had an impact, but the regime's strategy towards Eastern Sudan reiterates that realpolitik (such as securing oil export) took precedence over ideological factors.
The difference in partiality influenced the mechanisms underpinning cooperation. First, in Darfur, the government's partiality generated selective sanctions that contributed to violent communal conflicts by, lowering favoured groups' threshold for using violence, reducing the likelihood that disfavoured communities sought legislative solutions, and disrupting traditional conflict management. In Eastern Sudan, the regime's mainly impartial conduct facilitated effective sanctions and many disputes deescalated after punishments targeting perpetrators of intercommunal violence. Second, while the government's administrative adjustments (to shift power balances and to incite conflicts) created uncertain boundaries and aggravated the chances for co-existence in Darfur, boundaries remained largely undisrupted in Eastern Sudan. Third, as part of its partial agenda, the government endorsed violence in Darfur by undermining local actors and through distancing rules from local circumstances. In striking contrast, the government allowed prominent local actors to influence decisions in Eastern Sudan, which deescalated numerous disputes.
The empirical examination thus lent support to the three mechanisms. This does not mean, however, that the explanatory power of each mechanism was equally strong. For two of the mechanisms (sanctions and local rules), the comparison revealed causal connections between both the absence and presence of violent communal conflicts. The role of boundaries was less certain. While the mechanism contributed to violent communal conflicts in Darfur, the link to the absence of such conflicts in Eastern Sudan was not equally clear. Why? A conceivable reason for this is that while unclear boundaries might cause disputes, it is more difficult to figure out how they might influence how differences are managed. Disputes over boundaries repeatedly occurred in Eastern Sudan but did not escalate, indicating that how disputes are managed seems to be more important than the frequency of such differences. A key reason for why conflicts over boundaries were easier to settle in Eastern Sudan was that, in contrast to Darfur, borders were not generally drawn to shift power balances or to incite conflicts.
This study offers insights into the resolution and prevention of communal conflicts. Exposing mechanisms fostering intercommunal cohabitation facilitates consideration of procedures that can strengthen concurrence and insights on how to enhance community resilience to communal violence. Eastern Sudan demonstrates that communities can live in relative peace, despite extremely harsh conditions, if principles for cooperation remain intact. This includes a fair justice system where the severity of sanctions is in relation to the severity of the crime. Clear boundaries encourage cooperation so policies contributing to well-defined administrative units are likely to have a conflict-dampening effect. Furthermore, local ownership matters. To enhance intercommunal cooperation, local actors should be empowered and local circumstances carefully considered.
To what extent are these findings applicable outside Sudan? First, numerous countries suffering from violent communal conflicts share several characteristics-such as patrimonial structures, salient communal identification and severe land grievances-with Sudan. The article's findings are thus drawn from a relatively typical case. Second, the causal story attributes an essential importance to trust, which influences opportunities for cooperation and risks for conflict universally. Third, communal conflicts elsewhere, such as the severe Hema-Lendu conflict in the Democratic Republic of Congo illustrate interesting similarities to dynamics in this article. The dictator Mobutu Sese Seko favoured the Hema and this government bias 'poisoned relations between two communities that had previously coexisted, albeit uneasily, for many years' (Deibert, 2013, p. 2). Widespread impunity, or lack of appropriate sanctions, fuelled the fighting. Furthermore, boundaries was important. A land dispute sparked the fighting and administrative adjustments intensified it. Finally, extensive interference of neighbouring states (and the fact that policies from the international community did not reflect local realities) meant that adherence to local rules was minimal (Deibert, 2013). Although briefly sketching an intricate conflict leaves out key dynamics, it indicates that this article's theoretical argument has some bearing also on this context.

Disclosure Statement
No potential conflict of interest was reported by the author(s).