Protection or predation? Understanding the behavior of community-created self-defense militias during civil wars

ABSTRACT

During civil wars, some communities raise self-defense militias to protect themselves from insurgent predation, but these militias can end up mutating into predatory organizations. The extant literature has focused chiefly on the predatory propensity of state-created self-defense militias and has mostly overlooked why some community-created self-defense militias segue into predatory organizations while others eschew predation altogether. This study explains this phenomenon, drawing on in-depth interviews with active members of two community-created self-defense militias (Ahlu Sunna Waljama’a and Macawiisley) in Somalia. In doing so, two sequential mechanisms (sponsorship and mobility) that determine the propensity of predatory behavior are introduced. Self-defense militias that conduct offensive operations engage in predatory behavior, irrespective of whether they are sponsored locally or have external patrons. Externally sponsored self-defense militias that engage in offensive operations attract opportunistic recruits and become motivated by material benefits, while community-sponsored self-defense militias that conduct offensive operations instrumentalize their position to settle old scores against rival communities. By contrast, self-defense militias that restrict their operations to defensive activities typically recruit dedicated homegrown members, and are regulated by community-managed accountability mechanisms that prevent predatory and abusive behavior. This community control remains crucial for defensive self-defense militias, who must balance external patrons’ strategic aims with their local objectives.

KEYWORDS:

• Militias
• self-defense forces
• predation
• protection
• Somalia
• community-based armed groups
• Ahlu Sunna Waljama’a (ASWJ)
• Macawiisley

Introduction

While research on anti-insurgent militias has progressed significantly in recent years, the conditions under which community-created self-defense militias in fragile states engage in or avoid predatory violence have received little attention in civil war studies.¹ There is a growing view in civil war literature suggesting that self-defense militias formed by the state that engage in offensive counterinsurgency operations tend to turn predatory as a result of principal-agent problems stemming from diverging objectives and interests between states and self-defense militias. By contrast, community-created self-defense militias are usually static (i.e. with a geographically limited operational span) and less predatory due to community-managed accountability mechanisms.² There is an assumption that since community-created self-defense militias are formed spontaneously, they are driven by resistance against predatory insurgent groups and tend to have the least ‘proclivity toward abuse of local and rival communities’.³ However, anecdotal evidence suggests that some community-created self-defense militias segue into predatory organizations.⁴ This raises the question: under what conditions do community-created self-defense militias engage in or avoid predation?

This paper seeks to contribute to the emerging body of research on anti-insurgent self-defense militias in the civil war literature. In doing so, the aim is to provide theoretically and empirically grounded explanations – rooted in existing scholarship and empirical evidence from Somalia – for why some community-created self-defense militias might engage in or avoid predation. Moreover, this inquiry adopts the definition that self-defense militias always emerge as a community effort, with the sole aim of protecting the community from external violence, and that any state-created ‘self-defense’ militias are mostly recognized as paramilitary groups or referenced by the epithet ‘death squads’.⁵

Governments and local communities often sponsor self-defense militias to fill the capacity gaps of the state.⁶ Governments subcontract self-defense militias as a cheap force multiplier that allows them to tap into the local knowledge and power that a community group can wield against mutual enemies, such as insurgents and criminals, as well as shared political, ethnic, or religious rivals.⁷ Conversely, local communities sponsor self-defense militias to fill the political, social, and/or security gaps that emerge in the context of a state’s inability or unwillingness to safeguard communities’ interests against insurgent attacks. While self-defense militias are initially community-sponsored, many eventually become co-opted as auxiliary forces by the state, thereby becoming state-sponsored. The greater the extent to which they are armed by the state, the more autonomous and predatory they become. Pertinent examples include the Civil Defense Forces (CDF) in Sierra Leonne,⁸ which started as an umbrella organization of local protection groups but then acquired state sponsorship. The CDF was among the main perpetrators of violence against civilians during the country’s civil war.⁹ Similarly, the community-created Borno vigilantes in Nigeria were integrated into the government’s official counterinsurgency efforts¹⁰ and have been implicated in grave civilian abuses, including extortion, extrajudicial killings, and systematic sexual violence.¹¹ Similar observations have been noted in some community-sponsored self-defense militias that have been implicated in predatory and abusive behavior. For instance, self-defense militias in Mali (e.g. Dozo, Dan Na Ambassagou) have regularly engaged in serious abuses, including killings, pillaging, arson, and extortion.¹²

This paper identifies two sequential mechanisms that shape the way in which self-defense militias engage in or avoid predatory tactics. The first mechanism is sponsorship, which concerns whether a self-defense force becomes sponsored by an external patron (e.g. a state) or remains community sponsored. Akin to rebel groups, self-defense militias that acquire external patrons, such as politicians or governments, are more likely to attract opportunistic recruits and forego their initial community-oriented objective(s) in favor of a private agenda or a patron’s strategic aims.¹³ By contrast, those groups who lack an external patron rely primarily on community-contributed resources to sustain themselves and are driven by parochial aims that sustain their capacity for collective action, and tend to comprise dedicated homegrown recruits.¹⁴

The second mechanism is mobility. Self-defense militias that primarily engage in offensive operations (i.e. searching for insurgents across a given region), irrespective of the existence or type of sponsor, are more likely to be motivated by material pursuits or target rival communities than those without expansionary pretensions. Indeed, given the limited access to local support, the threshold for rapaciousness and dispensing predatory violence against civilians is considerably reduced for offensive self-defense militias that acquire external patronage. However, offensive self-defense militias that remain community-sponsored tend to engage in violent reprisals against rival communities, expel competitors for control over land and natural resources, and settle old scores.

By contrast, for self-defense militias that are primarily engaged in defensive operations, whether they turn to or refrain from predatory violence depends on their compliance with localized rules and norms. Defensive self-defense militias with local ties and community supervision are primarily interested in the security of their respective communities. Conversely, defensive self-defense militias with weak or no local ties and a concomitant lack of community control are more likely to become abusive towards their own and other communities.¹⁵

This paper assesses these two mechanisms by conducting a within-case comparison at ground level, using evidence from qualitative interviews with active and former members of two self-defense militias in Somalia. Since 2007, Somalia has been subjected to an undulating spiral of violence and suffering at the hands of the Al-Qaeda-aligned insurgent group, Harakat Al-Shabaab Al-Mujahideen (henceforth, Al-Shabaab). With the country’s security service in complete disarray and Somalia’s security forces unable to challenge the growing threat posed by Al-Shabaab,¹⁶ civilians are bearing the brunt of the conflict. Within the contours of Somalia’s civil conflict, several anti-Al-Shabaab self-defense militias have emerged as a reaction to the state’s incapability to provide security coupled with Al-Shabaab’s increasingly draconian measures, which have impacted the lives of ordinary civilians. One of these groups is Ahlu Sunna Waljama’a (‘People of the Sunna and Community’, ASWJ), and another is Ma’awiisley (Macawiisley in Somali, a reference to the sarong worn by many citizens). Both groups have similar origins – emerging as a response to Al-Shabaab’s attacks, racketeering, and extortion – and have been relatively successful in fighting and defeating Al-Shabaab in their respective territories, yet they differ in how they treat civilians. Macawiisley has provided protection sans predation, while ASWJ has become increasingly abusive towards civilians.

The remainder of this paper will proceed as follows. After setting the definitional scope of self-defense militias, the conceptual mechanisms of sponsorship and mobility are introduced. Drawing on in-depth interviews with ASWJ and Macawiisley members, the subsequent section will empirically discuss how the sequential mechanisms of sponsorship and mobility have affected ASWJ’s and Macawiisley’s propensity for predation. Finally, the paper considers the implications of its findings for scholars’ understanding of self-defense militias and suggests future research avenues.

Community-based armed groups

The corpus of works on self-defense militias remains limited, with the few existing studies examining this topic from within a wide range of disciplines, inter alia, criminal justice and criminology,¹⁷ political science,¹⁸ and anthropology.¹⁹ Some of the existing studies that look at community-based armed groups conflate self-defense militias with other types of non-state armed groups in fragile contexts (rebels, gangs, and paramilitary militias). Although self-defense militias share many similarities with paramilitary forces, their main differentiator is in militia–state relations. While paramilitary militias are created via a top–down state-driven approach (i.e. by the state),²⁰ self-defense militias come into existence through a bottom–up approach (i.e. from within the community).²¹

This study focuses on a particular type of non-state armed group, namely self-defense militias. Given the variety of definitions, understandings, and synonyms,²² on the one hand, and the interest in the instrumentalization of self-defense militias on the other, this study will adopt the broad definition of self-defense militia as an anti-insurgent armed group that is community-created, nominally pro-government, composed entirely of civilians, and has the primary purpose of protecting the local population from abusive insurgent and criminal groups. This definition shares similarities with those used in several other studies.²³

Furthermore, while self-defense militias have complex and often shifting motives and intentions, collective self-defense lies at the core of their identity and serves as a rationale for their violence. As local guardians seeking to fill gaps in state protections and provisions, the narratives of self-defense forces ‘include the defensive use of violence (in the name of a nation or for the survival of the community); the protection or liberation of communities; the resistance against injustices; and the defense of autonomy, self-determination, pride, or honor’.²⁴ Just as their motives are complex and varied, so too is their propensity for predatory behavior.

Sponsorship-mobility model

Why do some self-defense militias engage in large-scale predation while others avoid it altogether? Drawing on existing scholarship, this section will highlight and explore the theoretical underpinnings of why some self-defense militias engage in predatory behavior while others avoid it. To contribute to a better understanding of this phenomenon, this paper presents two intersecting and sequential mechanisms build upon the theoretical foundations of existing scholarship: sponsorship and mobility. These mechanisms are sequential in the sense that sponsorship precedes mobility in practice.

Sponsorship

Historically, communities have created militias to fill political, social, or security gaps that the state is incapable or unwilling to close. The bottom–up creation of such groups is typically spurred by several factors, including increasing incursions from insurgent groups, rising criminality, or increasing state repression.²⁵

Community-created self-defense militias are often co-opted by the state, thereby becoming state-sponsored.²⁶ Agents or elements within the state may come to sponsor a self-defense militia that was independently created by providing substantial funding, weapons, intelligence, expertise, and training. This type of sponsorship often occurs when self-defense militias and the state have mutual enemies; the state collaborates with these militias because they possess superior local knowledge and power in regions where insurgents and criminals operate against the interests of the militia’s constituency.²⁷ Examples of state-sponsored self-defense militias include the Anbar Awakening Movement in Iraq (also referred to as the Sons of Iraq), which were sponsored by the US military in order to fight against Al-Qaeda; the anti-Boko Haram Borno vigilantes were sponsored by the Nigerian government and institutionalized under the Civilian Joint Task Force (CJTF); the Kamajors of Sierra Leone, created to protect communities from the Revolutionary United Front (RUF) rebels and later sponsored by the Sierra Leonean government to fight against insurgents; and the Mexican Autodefensas which was an armed popular uprising against drug cartels, but which were later co-opted by the Mexican government and organized crime groups.²⁸

The sponsoring of self-defense organizations by external patrons, either state or otherwise, increases their propensity to become predatory.²⁹ Using data from the Pro-Government Militia (PGM) Database, Mitchell, Carey, and Butler found in a cross-country survey that the presence of government-aligned informal militias is associated with a significant increase in torture, disappearances, and killings.³⁰ Predation, when sponsored by a government, is associated with two related indicators. The first is that state authorities use loose, informal links to militias to evade accountability for human rights violations by publicly claiming the perpetrators are not under government control; the second indicator is that sponsored self-defense militias often break free of the control of their sponsor and become a liability to them, meaning that governments have difficulty controlling their sponsored militias, which can result in lower discipline and increased violence.³¹ Finally, researchers and organizations have documented empirical evidence related to externally sponsored self-defense militias that highlight gross human rights violations, criminal behavior, refusal to disarm, and actions that further weaken formal government structures and influence.³²

The nature of recruitment can also partly influence the propensity for predation, and several scholars have linked access to external resources with opportunistic recruits. For instance, Weinstein suggests that militias with early access to external resources can offer economic inducements to followers that enable them to recruit quickly, but at the cost of more predatory behavior toward civilians throughout the conflict.³³ This reinforces that some state-sponsored self-defense militias can be Janus-faced, serving simultaneously as sources of security and insecurity. Comparing different types of self-defense militias in Aceh, Barter concluded that state-sponsored self-defense militias tend to attract opportunistic recruits that target civilians, while those sponsored by the community to confront powerful rebels attract dedicated locals personally affected by the violence and are more likely to behave defensively.³⁴

Evidence supports the idea that self-defense militias linked to external patronage might attract extremist or opportunist members that are less concerned about reputational damage resulting from the use of violence³⁵ Examples of this include Yemen’s Popular Committees, which consisted of locals who banded together to resist and oust Jama’at Ansar al-Shari’a (AAS). However, after becoming sponsored by the Yemeni government and subsequently infiltrated by AAS members, they began engaging in looting and extorting cash at road checkpoints.³⁶ Similarly, the Mali government tacitly delegated several self-defense militias in Mali to fight against Al-Qaeda-linked insurgent groups. However, many turned abusive as they engaged in highly publicized mass atrocities and score-settling against rival ethnic groups.³⁷ These examples support the view that some state-sponsored self-defense militias might generally be protective of their communities but become predatory.³⁸

By contrast, self-defense militias that remain community-sponsored often retain their full capacity for collective action, are usually static, have higher morale, and are fundamentally motivated to protect their families and communities first; as a result, they tend to be less predatory. However, previous studies have almost exclusively focused on state-created ‘self-defense’ militias and scant attention has been paid to the proclivity of predation in community-created self-defense militias and under what conditions this occurs. Nevertheless, the period before a self-defense militia obtains external patronage might provide insights into the behavioral activities that deter them from engaging in repressive activities against civilians. One example is the Arrow Boys (AB) in South Sudan, a self-defense militia born out of the need to protect civilians from the Lords Resistance Army (LRA). The group achieved external recognition by controlling LRA attacks on the community and, according to many accounts, were more successful than various official military actors operating against LRA in the region. Before AB became sponsored by opposition groups³⁹ and subsequently turned repressive,⁴⁰ the group enjoyed widespread popular support from their community and rarely ventured beyond their towns and villages.⁴¹ Moreover, the group acted as both a protection and information hub for the local community, making them a ‘highly regarded group in the community’.⁴² Additionally, AB’s initial accountability to community decision-makers rather than an external patron, ‘regulated their behavior, resulting in very few abuses of power and helping to sustain popular support’.⁴³

Scholars have emphasized the link between strong community ties and lack of predation. For example, Tariq highlights that arbakai – Afghanistan’s grassroots self-defense militia – has abstained from predation largely because of its strong relations with the local community.⁴⁴ Similarly, Suarez argues that, to limit the predatory behavior by Mai-Mai self-defense militias in Congo, community members employed a combination of material support and verbal warnings to remind Mai-Mai of their primary obligation to protect the community.⁴⁵

Similarly, in the early 1990s in Liberia, factional rebel groups raided villages and urban domains, killing civilians; this prompted locals to organize self-defense militias, often along ethnic lines.⁴⁶ These self-defense militias often patrolled the streets at night, engaging in selective violence and filling what was essentially a security vacuum where the Liberian police were unable or unwilling to respond. As informal security providers closely connected to their local neighborhoods, they successfully reduced crime and largely shunned abusive conduct.⁴⁷ However, amid growing crime and violence in 2006 and given the state’s inability to deal with lawlessness, former Liberian Minister of Justice Frances Johnson Morris called on Liberians to form self-defense militias to augment the role of the formal security sector. The government’s legitimization of self-defense militias, enabled opportunists to infiltrate existing self-defense groups, with some morphing into abusive and quasi-criminal organizations engaged in armed robberies, extortion, and extrajudicial killings.⁴⁸ These examples support the view of several scholars who argue that self-defense militias that are firmly grounded in the local communities have fewer control problems and are more likely to act responsibly.⁴⁹

Mobility

Type of sponsorship alone is not a predictor of predation; it has to be accompanied by a mobility mode. Depending on a group’s sponsorship type, self-defense militias can pursue different mobilities (offensive or defensive operations) to sustain their anti-insurgent activities, leading them to adopt or abstain from predatory behavior. To conceptualize the operations of a self-defense militia, this paper defines mobility as a binary consisting of either (1) offensive operations, where self-defense militias seek out insurgents beyond their locality, and (2) defensive operations, where it remains static by focusing on securing its community from insurgent attacks. This paper argues that self-defense militias that conduct offensive operations (whether externally or community-sponsored) essentially turn predatory. However, whether self-defense militias with external patronage turn abusive or are restricted to protection is contingent on community oversight.

This paper argues that self-defense militias that conduct offensive operations are more likely to be predatory, and they are more likely to turn from defensive operations to offensive when they become state sponsored. Comparing the behavior of the ‘stationary bandit’ with that of the ‘roving bandit’, Olson finds that roving bandits tend to be more abusive than stationary ones, with little interest in the prosperity of their given community.⁵⁰ Externally sponsored self-defense militias often possess offensive capabilities and conduct attacks within the remit of counterinsurgency campaigns and engaging political rivalries. Yet there are several cases that highlight how quickly a state-sponsored defensive self-defense militia can adopt an offensive approach that leads to civilian victimization, particularly in the context of counterinsurgencies.⁵¹ An example of a group that altered their mobility is the Peruvian peasant self-defense force, rondas campesinas. They were formed initially by the community to protect against theft but were subsequently sponsored by the Peruvian government to fight against the Shining Path insurgency. Until 1989, the group’s modus operandi was entirely defensive and showed no signs of predatory behavior.⁵² It was only when they began engaging in offensive state-led counterinsurgency operations in 1989 that cases of human rights abuses were reported; for example, a 1993 Amnesty report identified disappearances and executions carried out by the rondas campesinas.⁵³

Similarly, from their inception as a self-defense militia up to becoming sponsored by an external patron, the South Sudanese AB group utilized an entirely defensive operational approach against LRA and rarely exhibited predatory behavior. Yet when they transitioned from community-sponsored to externally sponsored, they switched to offensive operations and became increasingly repressive against civilians, being accused of engaging in banditry, sexual slavery, and forced labor.⁵⁴

Although community-sponsored self-defense militias are initially defensive, by definition, many of these militias, adopt an offensive approach leading to abusive behavior, including violent reprisals against other rival communities, ostensibly to punish alleged support for insurgents, expel competitors to gain control over land and natural resources, and to settle old scores. For example, until 2015, the Dozo, a self-defense militia of traditional hunters in Mali, limited the scope of their operations to village protection. However, following the influx of armed Islamist groups in the region and Mali’s growing counterinsurgency campaign, the militia took the opportunity to engage in small-scale pogroms, banditry, pillaging, and large-scale livestock theft – all under the banner of ‘fighting’ against armed Islamists.⁵⁵ Similarly, a Dogon self-defense militia, Da Na Ambassagou (DNA), imposed a 29-day embargo on the Dogon village of Sogou after villagers refused to have a DNA camp in their village, forcing inhabitants to pay a ransom of 1,800 million CFA francs ($3,000).⁵⁶ In Somalia, two anti-insurgent clan militias, Habar Gidir and Biyomaal, staged internecine attacks to obtain control of the coastal town of Marka, leading to gross civilian abuses while justifying their offensive operations under the pretext of fighting against Al-Shabaab.⁵⁷

By contrast, those self-defense militias that rarely venture beyond their homes and villages, and so remain defensive in their operations, may or may not adopt predatory behavior, depending on the strength of community control. As mentioned earlier, self-defense militias’ raison d’etre is collective self-defense, which is what separates them from other violent non-state groups. Naturally, community-sponsored and defensive self-defense militias are bound by incentives that increase the rationality of abstaining from civilian targeting. For instance, following predatory behavior by different militias in Somalia’s capital Mogadishu in the early 2000s, communities mobilized and self-financed self-defense militias, named madani, for each neighborhood. These militias were accountable to the community elders and operated by arresting local criminals, responding to local distress calls, and chasing away roaming militias from other clans and neighborhoods.

This paper argues that whether defensive self-defense militias that acquire external patronage engage in abusive behavior or avoid predatory tendencies depends primarily on the strength of their ties to the local community, which regulates their behavior through community-managed accountability mechanisms. For example, in 2012, amid growing anti-Taliban sentiments, Andar communities staged a popular uprising comprised of different self-defense militias that ultimately evicted the Taliban insurgent group from Andar in Afghanistan.⁵⁸ State patrons rushed to support the counter-insurgents, leading to many of these varied groups becoming institutionalized under a government-sponsored paramilitary group unit, the Afghan Local Police (ALP). Highlighting the importance of community-controlled oversight mechanisms, Dearing finds that some units within the Andar ALP avoided predatory behavior because of local oversight that ensured compliance to localized rules and norms. Conversely, Andar ALP units that lacked this community accountability preyed on the local population.⁵⁹

A notable example of an externally sponsored defensive self-defense militias that avoided predation is the US-sponsored Sons of Iraq who were instrumental in clearing Al-Qaeda from their traditional territory and, because of strong oversight from local tribal elders, they almost entirely avoided predation.⁶⁰ However, this is not always the case. Following state sponsorship, defensive self-defense militias such as the CJTF in Nigeria,⁶¹ and the Popular Committees in Yemen,⁶² were infiltrated by opportunist recruits, had little concern for local accountability and control, and engaged in banditry and human rights violations.

In sum, the proposed model pathway (see Table 1) suggests the following: a community or externally sponsored self-defense militia that conducts offensive operations against insurgents is likely to turn predatory in pursuit of material benefits or to engage in score-settling against adjacent communities. By contrast, a community or externally sponsored and defensive self-defense militia will either adopt predatory behavior or be restricted to protection depending on the strength of community control.

Table 1. Sponsorship and mobility model of predation and protection

	Offensive	Defensive
External	Predation Examples: Civil Defense Force (Sierra Leone)^63 Imghad Tuareg Self-Defense Group and Allies (Mali)^64 Rondas Campesinas (Peru)^65	Predation Examples: Civilian Joint Task Force (Nigeria)^66 Popular Committees (Yemen)^67 Hezbi/Uprising (Afghanistan)^68 Protection Examples: Javanese and Gayo militias (Indonesia)^69 Anbar Awakening Movement (Iraq)^70 Nejat Council/Uprising (Afghanistan)^71
Community	Predation Examples: Dozo (Mali)^72 Dan Na Ambassagou (Mali)^73 Anti-Balaka (Central African Republic)^74	Protection Examples: Arbakai (Afghanistan)^75 The Rouga (Burkina Faso)^76 Madani (Somalia)^77

Methods

The case of Somalia is particularly useful for conducting an empirical test of the sponsorship and mobility model. Somalia has long been a byword for a failed state. Since the state collapse in 1991, the country witnessed a lateral growth in different types of militias that have become the hallmark of Somalia’s conflict landscape. These include factional militias, private business militias, jihadist militias, and, more recently, self-defense militias. With the exception of jihadist militias, all these militias are organized along clan-lines and created by local communities, often to shield themselves from insecurity, vulnerability, and contestation. Despite Somalia being awash with a plethora of militias, it is only in the last decade that the country has witnessed a rise of self-defense militias. Prior to that, most clan-based militias were formed to defend factional warlords and business leaders’ personal and business interests.

Since 2007 Somalia has been on a promising path of rebuilding the state; yet security remains elusive, with Al-Shabaab still controlling large swathes of the country. Clans started forming self-defense militias to undertake intelligence, security, and limited combat roles concerned with protecting their communities. The rise of self-defense militias in Somalia added a new dynamic to a conflict zone rife with extortion, predation, and land and resources theft. Presently, only two self-defense militias exist in the country that have actively engaged (and continue to engage) Al-Shabaab, namely ASWJ and Macawiisley. In short, Somalia presents an excellent controlled context to examine the behaviors of anti-Al-Shabaab self-defense militias longitudinally and uncover the sponsorship–mobility mechanisms over time.

This study draws its empirical findings from 64 in-depth telephone interviews conducted by the author between July and September 2019 with former and active members of the Macawiisley (n = 30) and ASWJ (n = 34) self-defense militias. Interviewees were primarily stationed in the Hiiraan, Middle Shabelle, and Galgaduud regions in Somalia. Regarding their ‘military’ ranks, one commander (taliye) and two deputy commanders (taliye-ku-xigeen) were interviewed for each group. The remainder of the interviewees were fighters. Each participant was selected based on their engagement within the group. All respondents were males aged between 20 and 60 years. The majority of participants from Macawiisley were pastoralists (xoolo dhaqato, n = 18), with farmers (beeraleey, n = 12) accounting for less than half. All respondents had been affected by Al-Shabaab-related violence and had participated in at least one skirmish against Al-Shabaab.

All telephone interviews with members of Macawiisley and ASJW self-defense militias were semi-structured and lasted between 40 and 90 minutes. Interviewees were asked to express their views about how their groups were formed, their reasons for joining, their ongoing motivations and challenges, their source(s) of support/funding, and the mode of their operations. Interviews were conducted in the Somali language and the data was coded as narrative analysis. Since all interviews were conducted under strict anonymity conditions, the interviewees’ names/identifiers were replaced with code names.

Empirical evidence from Somalia

Ahlu Sunna Waljama’a (ASJW)

Sponsorship

ASWJ was initially formed in late 1992 to protect traditional Sufi Islam from the encroaching jihadist groups that had begun gaining prominence in Somalia following the state’s collapse. Sufism has been a part of Somalia’s religious landscape since Islam first reached sub-Saharan Africa more than a millennium ago. During the 1990s and early 2000s, ASWJ’s opposition to extremist Islamist groups relied on non-violent tactics, such as religious preaching for the revitalization of Sufism and the unification of Sufi religious orders. The group, comprised mainly of civilians and religious scholars, was largely left alone by extremist Islamist groups and, until mid-2008, was reasonably inconsequential, lacking a political profile or military wing in Somalia’s theater of conflict.

Al-Shabaab (and its precursor, al-Itihaad al-Islaamiya) ignored ASWJ, or at least did not purposefully target the group. This changed in 2008 when Al-Shabaab’s chief military commander, Adan Hashi Ayrow, was killed by an American drone strike; following this, the group opened its doors to foreign fighters and began to enforce draconian edicts in areas under its control. Subsequently, Al-Shabaab expanded across the central regions and went on a spree of assassinating Sufi clerics,⁷⁸ banning Sufi practices, raiding mosques during Sufi rituals, and destroying Sufi shrines in central Somalia.⁷⁹ As Sufi civilian leaders began to feel increasingly targeted, they found themselves engulfed in the broader conflict; perceiving themselves as defenders of their religious practices and communities, they rose up against Al-Shabaab and created a self-defense militia.⁸⁰

Against the backdrop of the Ethiopian military’s withdrawal from Somalia in 2009 and Al-Shabaab’s mounting repression, traditional clans in the central region (Hawiye, Darod, and Dir) were affected by Al-Shabaab’s brutal tactics and served as a buttress for the ASWJ. Indeed, Somali clans in the region were growing increasingly disenfranchised by Al-Shabaab’s anti-freedom edicts. While civilians initially credited Al-Shabaab with restoring rule of law, banning their religious practices was a step too far.⁸¹

Al-Shabaab’s repressive measures led to ASWJ benefiting from an influx of local and dedicated volunteers to protect their Sufi traditions. The majority of ASWJ respondents interviewed in this study were part of the first wave of volunteers, and they emphasized that every recruit was in one way or another affected by Al-Shabaab’s violent acts and that their morale was exceptionally high.⁸² However, they also noted their inexperience, highlighting a lack of discipline and formal military training. To overcome this, the group relied on local former military tacticians who advised on military tactics .⁸³ The group was supported mainly through resources and transport provided by local clans; most of the arms came from the substantial government arsenals pillaged by warring clans at the onset of Somalia’s civil war.

ASWJ initially had trouble making strides against Al-Shabaab, but refused any external support. Following a failed initial incursion with 140 armed men against Al-Shabaab in early December 2008, ASWJ suffered a significant defeat in Guri’el. Ethiopia, which was in the mid of withdrawing its military forces from Somalia, temporarily postponed the process and signaled an intention to reinforce the besieged ASWJ.⁸⁴ However, ASWJ shunned any help save community support, not wanting to be perceived as an Ethiopian proxy. One ASWJ deputy commander remarked that, ‘Our struggle is based on popular local resistance against Al-Shabaab’s cruelty,’ and that ‘our God-given ability allowed us to reverse our first defeat against the Khawarij⁸⁵ and reclaim our home territories from them’.⁸⁶

In late December, however, ASWJ secured several momentous victories. The group captured the first town of contention, Guri’el, followed by a string of adjacent towns, including Adado and Galinsor. Residents who had initially fled Al-Shabaab’s oppressive rule and the ongoing battles between Al-Shabaab and ASWJ, returned to their hometowns and welcomed the new developments and victors. These events contributed significantly to ASWJ’s growing public support. An ASWJ fighter explained that the ‘jubilant welcome’ that was received ‘from our children, wives, parents, relatives, religious leaders, and clan elders was amazingly overwhelming’, and that ‘we showed the world and the Somali government that local communities could defeat the Khawarij [Al-Shabaab]’.⁸⁷ This shift thrust ASWJ into the political spotlight.

Following ASWJ’s initial victories against Al-Shabaab, many foreign and local actors were keen to sponsor the group’s fight against Al-Shabaab. Al-Shabaab initially leveled accusations against ASWJ of being an Ethiopian auxiliary force⁸⁸; in response, the then prime minister of Ethiopia, Meles Zenawi, held a press conference on January 152,009 – two weeks after ASWJ’s take-over of towns in central Somalia from Al-Shabaab – and remarked that Ethiopia was not ‘disappointed’ that Al-Shabaab was facing armed opposition from within Somalia and, while he emphasized that Ethiopia could not create such a group, he gestured that Ethiopia would support ‘such endeavor’.⁸⁹ Following the press conference, ASWJ reversed its early stance of refusing external help⁹⁰ and accepted Ethiopia’s sponsorship; in return, the group was supplied with assorted weapons and military training.⁹¹ Moreover, in the international arena, Ethiopia lobbied on behalf of ASWJ to acquire critical US funding for the group.⁹²

Alongside Ethiopia’s support, a host of opportunistic Somali politicians and warlords – many pro-Ethiopian – co-opted the Al-Shabaab struggle under the banner of ASWJ. This was the case for some notorious warlords and former members of the now-defunct CIA-funded anti-terrorism alliance, including Abdishukri Ali Hirsi, Abdi Waal, and Bashir Raage who, alongside their abusive clan-based militia, infiltrated the group and were appointed to senior positions.⁹³ As several decentralized ASWJ branches began to emerge in Somalia’s southern regions – more specifically in Hiiraan, Middle Shabelle, Gedo, and Bakool – several Somali politicians in those areas co-opted the militia from the outset.⁹⁴

The Transitional Federal Government of Somalia (TFG), however, had difficulty co-opting the group and was particularly apprehensive of ASWJ’s close relationship with Ethiopia, claiming that ‘ASWJ, with Ethiopia’s help, could become a competitor of the TFG and, potentially, a destabilizing factor in Somalia.’⁹⁵ Although ASWJ and the TFG signed several agreements of cooperation,⁹⁶ ASWJ had already cemented its status as a proxy group for Ethiopia in Somalia. This shaky relationship between ASWJ and successive Somali governments persisted until Ethiopia’s new leadership in 2018 gradually withdrew the state’s patronage from the militia. The group was demobilized and integrated into the Galmudug State Forces following a decisive battle in February 2020.⁹⁷

Mobility

ASWJ’s emergence as a self-defense militia was primarily built around defensive rather than offensive tactics. This defensive orientation was apparent in the group’s scope of operations in late 2008 and early 2009, where it sought to dislodge Al-Shabaab from its home territories. Key cities in Galgaduud were the first to be removed from Al-Shabaab’s grip, including Guri’el, Herale, Abudwaq, DhusaMareb, Masagawa, and Adado – cities primarily inhabited by the clans that formed the bulk of ASWJ.

However, after accessing Ethiopia’s patronage and being subsequently co-opted by pro-Ethiopian politicians and warlords, the group became increasingly ambitious in its political endeavors and switched from defensive to offensive tactics against Al-Shabaab, and other pro-government actors that stood in their way. This was most visible in late 2009, when ASWJ collaborated with Somali politicians and warlords to establish decentralized branches beyond its traditional stronghold of Galgaduud, to Hiiraan, Middle Shabelle, Bakool, and Gedo. For example, warlord-turned-Islamist Abdishukri, who co-opted the movement in 2010, utilized the ASWJ militia to overthrow the government-aligned Himan and Heeb administration in Adado in October 2010, a move that generated a strong rebuke from the Somali government.⁹⁸ Moreover, to fulfill its political aspirations, ASWJ challenged the Somali government and conducted several offensive operations to prevent the government administration from taking root in areas under ASWJ control.⁹⁹

Following ASWJ’s acquisition of patronage and departure from its initial focus on protecting communities toward offensive, politically-motivated operations, strong accusations of human rights abuses and predatory behavior were leveled against the group, including child recruitment, random shootings of civilians, extrajudicial killings, and extortion.¹⁰⁰ An ASWJ commander remarked that, ‘Since we received external backing from Ethiopia, many politicians and militia leaders joined ASWJ or used our name, and this damaged our credibility considerably and reinforced the message that we are a callbox’.¹⁰¹ In response to the predatory behavior of some members of ASWJ, the commander stated that, ‘The actions of many of our members against civilians went against our founding principles, and many got away with it, depending on their connections to some of the senior people’. He further added that, ‘because of some of our members’ Mooryaan behavior,¹⁰² it damaged our credibility as a respectful and honorable group and we lost community support.’¹⁰³

Researchers and human rights organizations have empirically documented abuses against civilians committed by ASJW. In late 2009, amid its offensive push, the group recruited hundreds of children, both voluntarily and forcibly.¹⁰⁴ It also engaged in a series of extrajudicial and summary killings, having tortured and killed scores of individuals accused of being pro-Al-Shabaab.¹⁰⁵ The group was also accused of opening indiscriminate fire at civilians.¹⁰⁶ Civilians in ASWJ-controlled areas complained of excessive ‘taxation’, with frequent shooting at those who refused to pass through the group’s tax-collection checkpoints.¹⁰⁷ This abusive behavior cemented the shift in ASWJ’s focus from protection to engaging in predatory behavior.

Macawiisley

Sponsorship

The Macawiisley self-defense militia was initially formed by civilians in 2014 to respond to Al-Shabaab’s excessive ‘taxes’, humiliation, and witnessing their children being forcibly recruited into Al-Shabaab’s armed machinery.¹⁰⁸ Al-Shabaab extracted disproportionately high levies on harvests and zakat (Islamic alms) from pastoralists and farmers. Indeed, multiple Macawiisley respondents in Hiiraan described how Al-Shabaab forcibly took hundreds of their camels as ‘zakat’ payment and attempted to sell them in the neighboring region Galgaduud. However, given that the camels bore clan brand marks, no local clan in Galgaduud was willing to buy them, as this would inevitably lead to a clan conflict; instead, Al-Shabaab slaughtered the camels and sold the meat.¹⁰⁹ Similarly, a Macawiisley commander explained that they ‘used to heavily tax us at least four times a year, especially during Islamic holidays, and told us that it was an Islamic obligation and the raised money would go toward feeding the poor, except that it never went to the poor’.¹¹⁰

Operating in parallel with taxation was forced recruitment. Al-Shabaab abducted children as young as eight and either sent them to their training facilities or, depending on the level of fighter scarcity, thrust them to the frontline.¹¹¹ The majority of the respondents in this study highlighted that they had lost at least one child to Al-Shabaab.¹¹² A deputy commander recounted that he reached a ‘breaking point’ when his eight-year-old son began ‘crying and screaming’ and that ‘he begged me to save him’.¹¹³ In addition to their children being forcibly recruited, many of the respondents were forced by Al-Shabaab to send their wives to an undisclosed location, every Tuesday and Saturday.¹¹⁴

Unable to endure Al-Shabaab’s callous tactics, dozens of villagers exchanged some of their possessions for arms and ammunition and vowed to liberate their homes from Al-Shabaab, calling themselves Macawiisley. Their remaining possessions, mainly livestock and farms, were violently usurped by Al-Shabaab as a punishment for creating a self-defense militia. Initial material support was not always forthcoming from the local populations and, in the absence of palpable victories against Al-Shabaab, Macawiisley relied primarily on dedicated homegrown fighters.

The majority of the early Macawiisley fighters were pastoralists, comprising men above 40 with no previous military training. One Macawiisley fighter stated that, ‘Everybody knew everybody and we were all related to each other in some way or another, whether through family, marriage or being neighbors’ and that, ‘because of the existing trust and our shared experiences, we were highly motivated to fight for what was right’.¹¹⁵ Multiple Macawiisley respondents revealed that, given the heightened fear of Al-Shabaab in their home villages, civilians there did not want to imperil themselves by providing Macawiisley with material support, but that changed when they ‘started recovering villages from Al-Shabaab’.¹¹⁶ These successes were the catalyst for increased public support in the Hiiraan and Middle Shabelle regions and the addition of fresh, eager recruits of villagers directly affected by Al-Shabaab’s rule.

However, ordinary civilians paid a hefty price for providing support to Macawiisley. At times, Al-Shabaab tried to curtail the growing popular support for Macawiisley by engaging in summary executions and mass kidnappings/imprisonment of civilians that supported Macawiisley,¹¹⁷ and staged civilian protests denouncing the militia.¹¹⁸ As recently as February 2020, Al-Shabaab burned homes in Yaaq Bariweyne and several smaller villages, for supporting a new branch of Macawiisley in Lower Shabelle.¹¹⁹ In response, Macawiisley conducted regular prison raids to free their civilian supporters.¹²⁰

Although Macawiisley, from its inception, was informally aligned to the government, state sponsorship was not directly forthcoming. The Somali government somewhat tolerated the militia, providing scant resources (i.e. food supplies), but official state support was mostly absent. Knowing that they could not expand without strong state support, Macawiisley made several pleas for official support to sustain their anti-Al-Shabaab operations, but felt it was entirely ignored by the Somali government.¹²¹ The group’s members emphasized this lack of support from the government (and the international community) in interviews conducted by Somali media outlets.¹²² Several Somali parliamentarians who visited the militia (as part of photo opportunities and to claim credit for their achievements) publicly criticized the government’s rebuff of the group, stressing that Macawiisley relied on community support and would be unable to expand its operations in the long run.¹²³

Despite the militia’s relative success in its fight against Al-Shabaab, the Somali government has refused to sponsor the group and provide critical state assistance, a decision arguably rooted in past historical experiences. Sponsored militias have played a critical role in Somalia’s chronic insecurity for the past 30 years and have proven to be a barrier on Somalia’s road to stability and state-building.¹²⁴ Similar to ASWJ, state sponsorship of Macawiisley could potentially incentivize other communities to form self-defense militias, seek state sponsorship, and set the stage for opportunistic recruits and co-option by Somalia’s political actors.

Mobility

Since Macawiisley’s formation in 2014, the group has not strayed beyond self-defense operations against Al-Shabaab, and always within the confines of their home territory. All Macawiisley respondents emphasized that they consider where their clan members live to be their home territory and boundaries.¹²⁵ Indeed, Somalis are mainly transhumant pastoralists and have traditionally established indicative border lines that define the contours of their imagined clan boundaries. Carrying out attacks against Al-Shabaab in another clan’s territory is considered an act of trespass, leading to inter-clan conflict.

As such, localized Macawiisley militias have restricted the scope of their territory to their respective clan zones. For instance, Hawadle and Gaalje’el Macawiisley confine their defensive operations to the Hiiraan region, the clans’ stronghold, while the Abgal Macawiisley restricts their sphere of defensive operations to Middle Shabelle, and the recently created Rahanweyn Macawiisley limits their defensive operations to the Bakool region.¹²⁶ A Macawiisley commander in Hiiraan explained that they ‘do not have any communication with the Macawiisley defending their home territory in Middle Shabelle’, but that they were aware ‘that they exist and that they are defending their families and livestock from Al-Shabaab, but they do not enter our areas and we do not enter theirs’.¹²⁷ Similarly, a deputy Macawiisley commander in Middle Shabelle remarked that their mission is to defend their families and community elders from Al-Shabaab; such, they answer only to them. He further added that, ‘Since this is our home, we know the area inside out, whereas if we supported the Macawiisley in other regions, we would be useless to them as we do not know their territory’.¹²⁸

As a result of adopting a defensive approach, Macawiisley devised a unique security strategy whereby Macawiisley farmers were tasked with defending the villages while pastoralists conducted defensive attacks from the periphery of their clan’s borders.¹²⁹

The pastoralist members of Macawiisley often walk for hours actively seeking out Al-Shabaab units that have entered their clan territory and directly engage or ambush them. For instance, on May 28 2015, two Macawiisley units left Buloburde town in Hiiraan and walked to Al-Shabaab-occupied villages, with one unit attacking Buurweyne village (30 km northwest of Buloburde), and the second unit attacking Af’ad (31 km southeast of Buloburde). They briefly captured the villages, stole guns from fleeing Al-Shabaab members, and then returned to Buloburde.¹³⁰ The group was equally successful in liberating their home villages from Al-Shabaab.¹³¹

Macawiisley farmers, on the other hand, formed the majority of the defense against incoming Al-Shabaab aggression. For instance, in October 2014, Al-Shabaab occupied a farming village near the Mahaaday district in Middle Shabelle and, as part of their racketeering scheme, levied fees on sales of agriculture produce in the town. Macawiisley farmer villagers launched an attack that resulted in Al-Shabaab vacating the village.¹³² A similar occurrence was observed in Hiiraan, where Al-Shabaab instigated an attack on farming villages between Buloburde and Jalaqsi in November 2014, but failed to make inroads because of Macawiisley farmers.¹³³

The unique defensive core–periphery strategy adopted by the Macawiisley pastoralists and farmers, coupled with the organic community support and oversight, has enabled them to stick to their founding objective of protecting the community from Al-Shabaab and, thus, avoid predation. Multiple Macawiisley respondents emphasized that their fighters are bound by local accountability, primarily from community elders, that ensures they avoid abusive behavior because ‘attacking civilians is like attacking your family relatives’.¹³⁴ Virtually all of their equipment, food, and legitimacy are derived from their respective local communities.¹³⁵ This observation was confirmed by third-party actors, including Somali military commanders and diaspora members, who were impressed with the militia’s lack of civilian targeting.¹³⁶ Much like the proverbial incentive not to ‘bite the hand that feeds you’, Macawiisley members were acutely aware of losing popular support if they abandoned local customs and norms and exploited the communities that had morally and physically supported them.

Looking forward

This paper advances our understanding of why anti-insurgent self-defense militias might engage in or abstain from abusive behavior against civilians. The few studies that have been published on self-defense militias have primarily examined the predation of self-defense militias in the context of state-created militias and sponsored proxies. This paper builds upon the existing literature of civil war studies exploring sources of heterogeneity among civil war combatants¹³⁷ by producing a model that examines predation from the perspective of community-created self-defense militias that either remain sponsored by their local communities or acquire external patronage. This paper argues that self-defense militias’ propensity to engage in or abstain from predation rests on two sequential mechanisms: sponsorship (whether they are externally or community-sponsored) and mobility (whether they conduct offensive or defensive operations).

Self-defense militias that are externally sponsored and conduct offensive operations against rebels tend to attract opportunistic and abusive recruits, have limited access to local support networks, and so forego their founding local objectives in place of self-interest or their patron’s strategic aims. Moreover, as the accountability transfers from the community to an external patron and the group embarks on offensive operations, they begin to operate with diminished community oversight and turn increasingly repressive. On the other hand, community-sponsored self-defense militias that engage in offensive operations will often victimize rival communities under the guise of fighting against insurgents.

By contrast, defensive self-defense militias that are either externally or community-sponsored avoid predation if they are regulated by their local communities and remain true to their founding objective of providing local protection, attracting dedicated fighters from their constituencies. As they remain accountable to the community they are protecting, they maintain their popular support and generally avoid predation. However, without this community-managed accountability mechanism, these defensive self-defense militias often prey on the communities they are supposed to protect.

Drawing empirical insights from unique interview data with active members of ASWJ and Macawiisley self-defense militias in Somalia, this paper provides a comparative example of how two community-created self-defense militias that emerged to fight against Al-Shabaab insurgency have exhibited divergent behavior towards civilians, with ASWJ displaying predatory tactics by engaging in abusive behavior toward civilians while Macawiisley has abstained from predation.

Upon its initiation, ASWJ took up arms against Al-Shabaab following the desecration of Sufi gravesites and the killing of Sufi clerics by Al-Shabaab. Although a grassroots self-defense militia, the group later acquired sponsorship from Ethiopia and was co-opted by several Somali warlords and politicians and their abusive militias. With this influx in its ranks, and to expand ASWJ’s political influence, the group shifted its modus operandi from defensive to offensive operations. They attacked and occupied non-Al-Shabaab-controlled towns, exacerbating clan divisions. This encouraged predatory behavior and the group engaged in forced child recruitment, random shootings of civilians, extrajudicial killings, and extortion.

Macawiisley, on the other hand, took up arms after growing impatient with Al-Shabaab’s incessant ‘taxation’ and child recruitment in Al-Shabaab-held areas. Although Macawiisley actively sought external assistance, neither the Somali government nor any other external patron ever officially sponsored the group. This forced them to rely on community support and contributions and to recruit dedicated homegrown fighters affected by Al-Shabaab’s violence. Past research has found that militias that rely on existing social networks for recruitment are less likely to attack civilians.¹³⁸ As a result of maintaining their original objectives and being accountable to communities, the group has never ventured beyond its clan territories. Instead, it devised a unique strategy wherein Macawiisley farmers form the central defense of villages under their control, while the Macawiisley pastoralists conduct pre-emptive defensive operations at the periphery of their clan territory. This cooperative strategy strengthened the militia’s popular support, forced Al-Shabaab to vacate Macawiisley’s strongholds, and prevented the group from venturing into rival clans’ territories.

However, the civilian support for Macawiisley has been tested by indiscriminate retributive violence by Al-Shabaab against pro-Macawiisley communities, including the burning and looting of entire villages, mass kidnappings, and summary executions by Al-Shabaab. A similar finding was reported by Clayton and Thomas who argued that while self-defense militias undermine civilian support for insurgent groups, their popularity can invite increased insurgent violence against civilians, thus prolonging civil conflict.¹³⁹

Somalia is finding it extremely difficult to regulate existing self-defense militias and struggles with the continual formation of new groups. In the past two years alone, two anti-insurgent self-defense militias have emerged, Ali Qaran in the Bakool region and Shabeeloow in the Mudug region.¹⁴⁰ The findings of this paper will inform a greater understanding of why some clan-based militias in fragile settings, such as Somalia, turn into predatory organizations while others reject predation. Indeed, Somalia has a tumultuous history with clan-based militias, the majority of which have emerged as grassroots projects. For instance, Somali pirates were initially a self-defense militia comprised of fishers fighting against illegal foreign fishing vessels off the Somali coast. However, as they acquired external patronage and switched their mobility to offensive operations, they transformed into armed robbers at sea, operating far beyond the coast of Somalia and into the Indian Ocean.¹⁴¹ In sum, since most state-sponsored self-defense militias have a higher proclivity to ‘turn bad’, state sponsorship of these groups would inevitably obstruct, and potentially reverse Somalia’s recent strides on its path to stability.

Furthermore, important lessons can be drawn from the model for Somalia’s current counterinsurgency against Al-Shabaab. The literature emphasizes the benefits of counterinsurgent self-defense militias in defeating insurgents.¹⁴² As such, this paper’s findings suggest that strong community accountability and strictly defensive operations can prevent predatory behavior in self-defense militias. Thus, the Somali government can play an essential role in empowering the community-led accountability mechanism acting on the self-defense militias that continue to emerge in Al-Shabaab-held areas. This way, such self-defense militias can actively contribute to defensive counter-insurgent operations while their proclivity to engage in predatory behavior will remain limited. Some studies suggest that when insurgent groups target communities, it can lead to explosive growth in militias, which may contribute to the decisive defeat of insurgent groups.¹⁴³ However, when these same self-defense militias turn predatory and indiscriminate, this can contribute to a proliferation of militias and fan the growth of insurgent groups, leading to prolonged conflicts.¹⁴⁴

The sequential mechanisms presented in this study might offer a greater understanding of why community-created self-defense militias engage in or avoid predation in the context of armed conflict, and how this could potentially apply to other cases. The context of Somalia provides useful insights related to self-defense militias fighting against jihadist insurgencies. This is relevant as Africa has overtaken the Middle East as the jihadist insurgent battleground and has witnessed a proliferation of self-defense militias in fragile states, particularly in the Sahel region.¹⁴⁵ Consequently, weak state actors have decided to sponsor and tacitly outsource to several self-defense militias in their fight against extremist actors. Not only does this fuel local conflicts, as many of these self-defense militias have adopted an offensive approach and comprise opportunistic members, but many were also crystallized along identity lines. Following the occurrence of large-scale human rights abuses by some of these self-defense militias, governments have struggled to disarm them, leading to growing anti-government sentiments and further exacerbating pre-existing fault lines.¹⁴⁶ Finally, future research could explore how the different types of sponsorship and mobility could affect the propensity for predatory behavior among other types of anti-insurgent militias in civil wars.

Disclosure statement

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

Additional information

Notes on contributors

Mohammed Ibrahim Shire

Dr Mohammed Ibrahim Shire is a senior lecturer in Security and Risk at the University of Portsmouth. He specialises in the increasingly complex world of risk and security and is multidisciplinary by training and experience. His research agenda takes a broad interdisciplinary approach to terrorism and political violence in Africa and crosscuts aspects of criminology, political science, peace and conflict studies and computer science.

Notes

1. Peic, “Civilian Defense Forces“; Eck, “Repression by Proxy.”

2. Carey and Mitchell, “Progovernment Militias,” 135–137.

3. Felbab-Brown, “Hurray for Militias? ” 259.

4. For instance, Anti-Balaka, a predominantly Christian militia in Central African Republic emerged as an organic expression of community needs for self-defense against predation by Seleka rebels. However, as the conflict progressed, the group carried out mass killings, engaged in looting and destruction of properties, mostly against Muslim communities but also Christian civilians; Kokpakpa, “U.N. Official Says “terrifying” Level of Hatred in Central African Republic.”; Schuberth, “The Challenge of Community-Based Armed Groups. ”

5. Raleigh, “Pragmatic and Promiscuous, ” 287.

6. Peic, “Civilian Defense Forces. ”

7. Eck, “Repression by Proxy, ” 928–930.

8. Hoffman, “The Meaning of a Militia, ” 641–643.

9. Keen, Conflict and Collusion in Sierra Leone, 268.

10. The Economist, “Nigerian Vigilantes – The Home Guard. ”

11. Crisis Group, “Double-Edged Sword, ” 20.

12. HRW, “Mali: Militias Kill Over 75 Civilians. ”

13. Mitchell, Carey, and Butler, “The Impact of Pro-Government Militias, ” 818.

14. Weinstein, Inside Rebellion.

15. Dearing, “A Double-Edged Sword, ” 589–592.

16. Ibrahim Shire and Hersi, “Brothers in Arms, ” 13–15.

17. See, for example, Johnston, “What Is Vigilantism? ”; Haas, de Keijser, and Bruinsma, “Public Support for Vigilantism. ”

18. See, for example, Jentzsch, Kalyvas, and Schubiger, “Militias in Civil Wars, ” 2015; Estancona et al., “Civilian Self-Defense Militias in Civil War”; Mazzei, Death Squads or Self-Defense Forces?

19. See, for examplem McGee, “Power, Violence, Citizenship and Agency. ”

20. Examples are the Autodefensas Unidas de Colombia (AUC) and Janjaweed.

21. Kan, The Global Challenge of Militias, 29–30.

22. These include vigilantes, civilian defense forces, protection millitias, civilian self-defense, and civil defense forces.

23. Clayton and Thomson, “Civilianizing Civil Conflict, ” 499; Schuberth, “The Challenge of Community-Based Armed Groups, ” 300–304.

24. Schneckener, “Militias and the Politics of Legitimacy, ” 804.

25. Schuberth, “The Challenge of Community-Based Armed Groups, ” 301–304.

26. Staniland, “Militias, Ideology, and the State, ” 774–775; Jentzsch, Kalyvas, and Schubiger, “Militias in Civil Wars, ” 756.

27. Hughes and Tripodi, “Anatomy of a Surrogate”; Kilcullen, “counterinsurgency Redux, ” 121–122.

28. Phillips, “Inequality and the Emergence of Vigilante Organizations, ” 1361.

29. Barter, “State Proxy or Security Dilemma? ”; Kan, The Global Challenge of Militias, 41–43.

30. Mitchell, Carey, and Butler, “The Impact of Pro-Government Militias. ”

31. Carey, Colaresi, and Mitchell, “Governments, Informal Links to Militias”; Mitchell, Carey, and Butler, “The Impact of Pro-Government Militias. ”

32. Dufka, “Self-Defense Group Abuses in Central Mali, ” 89–95; Mitchell, Carey, and Butler, “The Impact of Pro-Government Militias, ” 825–830.

33. Weinstein, Inside Rebellion, 7.

34. Barter, “State Proxy or Security Dilemma? ”, 76.

35. Kaldor, “New & Old Wars, ” 71–93; Ron, “Territoriality and Plausible Deniability, ” 297.

36. Gaston and Dawsari, “Waiting for Change, ” 85.

37. Ross, “Mali Struggles to Disarm.”

38. Andersen, Møller, and Stepputat, Fragile States and Insecure People?, 67–122.

39. Boswell, “Spreading Fallout, ” 10–11.

40. Mccrone, “Hollow Promises, ” 7–8.

41. Koos, “Why and How Civil Defense Militias Emerge, ” 1047–1048.

42. CSG, “Non-State Security Providers, ” 19.

43. Schomerus and Taban, “Arrow Boys, Armed Groups and the SPLA,” 16–17.

44. Tariq, “Tribal Security System (Arbakai), ” 3–8.

45. Suarez, ‘“Living between Two Lions”, 62.

46. UN Secretary-General, “Sixteenth Progress, ” 6–7.

47. McCall, “Liberian Civilians Take Law. ”

48. Corey-Boulet, “Vigilante Groups in Liberia”; AllAfrica, “Vigilante Linked to Horrific Murder. ”

49. Ahram, “Learning to Live with Militias”; Clayton and Thomson, “The Enemy of My Enemy Is My Friend, ” 930; Barter, “State Proxy or Security Dilemma? ” 83–88; Peic, “Civilian Defense Forces, ” 163; Clayton and Thomson, “Civilianizing Civil Conflict, ” 501.

50. Olson, “Dictatorship, Democracy, and Development, ” 568–569.

51. Romero, “Reform and Reaction, ” 178; Harnischfeger, “The Bakassi Boys, ” 41–42.

52. Fumerton, “Rondas Campesinas in the Peruvian Civil War, ” 487.

53. Amnesty International, “Ten Settlers Massacred. ”

54. Boezio, “Special Representative Patten Welcomes the Release of Abducted Women. ”

55. Dufka, “Self-Defense Group Abuses in Central Mali. ”

56. Crisis Group, “Reversing Central Mali’s Descent. ”

57. SomaliFanTv, “Habargidir Oo Biyomaal Ku Edeeysay. ”

58. Bezhan, “Afghan Villagers Hit Back, ”.

59. Dearing, “A Double-Edged Sword, ” 592–597.

60. Clayton and Thomson, “The Enemy of My Enemy Is My Friend, ” 924–925.

61. Nagarajan, “Protecting and Harming Civilians, ” 17–20.

62. Gaston, “Sustainable Security in Yemen. ”

63. See note 9 above, 268.

64. UN Secretary-General, “Report of the Secretary-General, ” 4–5.

65. See note 53 above.

66. Nagarajan, “Protecting and Harming Civilians. ”

67. Gaston and Dawsari, “Waiting for Change, ” 22–23.

68. See note 59 above, 597.

69. Barter, “State Proxy or Security Dilemma?” 76.

70. Wilbanks and Karsh, “How the “Sons of Iraq” Stabilized Iraq,” 58–62.

71. Dearing, “A Double-Edged Sword, ” 597–601.

72. See note 55 above.

73. Ibid.

74. Kokpakpa, “Level of Hatred in Central African Republic. ”

75. See note 44 above, 8.

76. Crisis Group, “Burkina Faso. ”

77. Le Sage, “Stateless Justice in Somalia, ” 49–50.

78. Xasan, “Wadaad Caan Ah Oo Xalay.’

79. HRW, “Harsh War, Harsh Peace Abuses by Al-Shabaab, ” 26–27.

80. Domenicantonio, “With God on Our Side, ” 66.

81. Hassan, “Inside Look at the Fighting, ” 2.

82. Interviews with ASWJ members, Galgaduud, July 2019.

83. Ibid. 2019.

84. Mohamed, “Somalia: Ethiopian Troops Said Heading. ”

85. ‘Khawarij’ is a label that ASWJ uses to denigrate Al-Shabaab, implying that they are a continuation of the extremist sect of Al-Khawarij that broke away from mainstream Islam hundreds of years ago.

86. Interview with ASWJ deputy commander, Galgaduud, July 2019.

87. Interview with ASWJ fighter, Galgaduud, July 2019.

88. See note 81 above, 2.

89. Addis Fortune, “Ethiopia: Status Quo, Patience, Time Not Against Us. ”

90. See note 82 above, 2019.

91. CISAC, “Ahlu Sunna Wal Jama. ”

92. Berhane, “Ethiopia’s “eccentric, Elusive” Spy Chief with US Ambassador. ”

93. Mareeg, “Isbaheysigii Argagixiso Ladirirka. ”; Goobjooge, “Ahlu Suna Oo Cabdi Waal. ”

94. Mareeg, “Mas’uul Ka Tirsanaa Maamulkii G/Hiiraan. ”

95. See note 80 above, 66.

96. Africa Research Bulletin, “Somalia: Pact With Ahlu Sunna, ” 8315.

97. Osman, “Somalia’s Sufi Muslim Leaders Surrender. ”

98. SomaliTalk, “Sheikh Shariif Oo Cambaareeyey Weerarkii. ”

99. Mareeg, “Dhuusamareeb Iyo Guriceel Oo u Kala Dhacay. ”

100. Day, Felbab-Brown, and Haddad, “Hybrid Conflict, Hybrid Peace,’ 141.

101. ‘Callbox’ is disparaging term used by ordinary Somalis to denote political stooges that accept anything for a couple of coins.

102. ‘Mooryaan’ is a derogatory term that carries connotations of thuggish and criminal behavior

103. Interview with ASWJ commander, Galgaduud, July 2019.

104. United Nations, “Report of the Secretary-General, ” 6.

105. Diini, “Ciidammo Ka Tirsan Ahlu-Sunna Waljamaaca”; NewSomali, “Fadeexadaha Ciidamada Dowladda”; HRW, “Somalia: Pro-Government Militias Executing Civilians. ”

106. HRW, “Somalia: End War Crimes. ”

107. Mudug24, “Maamulka Ahlu-Sunna Oo Canshuur. ”

108. Faruk and Bearak, ‘“If I Don’t Pay, They Kill Me”’; HRW, “Somalia: Al-Shabab Demanding Children. ”

109. Interviews with Macawiisley members, Hiiraan, August 2019.

110. Interview with Macawiisley commander, Middle Shabelle, August 2019.

111. Burke, “Al-Shabaab Plundering Starving Somali Villages”; HRW, “Somalia: Al-Shabab Demanding Children. ”

112. Interviews with Macawiisley members, Hiiraan and Middle Shabelle, August 2019.

113. Interview with Macawiisley deputy commander, Middle Shabelle, August 2019.

114. See note 112 above, 2019.

115. Interview with Macawiisley fighter, Middle Shabelle, August 2019.

116. See note 112 above, 2019.

117. Dalsan, “Alshabaab Abducts 30 People Near Adale. ”

118. Wardoon, “Shabaabka Oo Billaabay. ”

119. Yuusuf, “Al-Shabaab Oo Deegaanka Yaaq-Bari Weyne. ”

120. Shaaciye, “Kooxda Macaawiisley Oo Sheegtay Inay Siidaayeen. ”

121. See note 112 above, 2019.

122. Dal Jire, “Maxa Kataqaan Ciidanka Macawisley”; SNTV, “Dadka Degan Tuulooyinka Shabeelada Dhexe. ”

123. Dalsan, “Xil Cabdi Shire Oo Si Kulul”; Somali Cable, “Xildhibaan Daahir Jeesow. ”

124. Menkhaus, “Governance without Government in Somalia, ” 88; Day, Felbab-Brown, and Haddad, “Hybrid Conflict, Hybrid Peace, ” 126–127.

125. See note 112 above, 2019.

126. BBC, “Maxaad Ka Taqaanaa Maleeshiyada Cali Qaran? ”

127. Interview with Macawiisley commander, Hiiraan, August 2019.

128. Interview with deputy Macawiisley commander, Middle Shabelle, August 2019.

129. See note 112 above, 2019.

130. See note 109 above, 2019.

131. Cumar, “Horjoogayaal Iyo Maleeshiyaad Ka Tirsanaa Al Shabaab. ”

132. Radio Muqdisho, “Dad Beeraley Ah Oo Maleeshiyaad. ”

133. Diini, “Dagaal Al-Shabaab Iyo Dad Shacab Ah. ”

134. See note 112 above, 2019.

135. Ibid. 2019.

136. CBA TV, “DFS Oo Shaacisay. ”

137. Barter, “State Proxy or Security Dilemma? ”; Weinstein, Inside Rebellion; Olson, “Dictatorship, Democracy, and Development”; Mitchell, Carey, and Butler, “The Impact of Pro-Government Militias. ”

138. Forney, “Who Can We Trust with a Gun? ” 825; Humphreys and Weinstein, “Handling and Manhandling Civilians, ” 431; Stanton, “Regulating Militias, ” 901.

139. Clayton and Thomson, “Civilianizing Civil Conflict, ” 499.

140. VOA, “Dagaal Culus Oo Dhex Maray Dad Shacab Ah Iyo Al-Shabaab. ”

141. House of Commons Foreign Affairs Committee, Piracy off the Coast of Somalia, 13.

142. Cline, “Sharing the Load. ”

143. Degregori, “Harvesting Storms. ”

144. Aliyev, “Pro-Regime Militias and Civil War Duration, ” 631.

145. BBC, “Is Africa Overtaking the Middle East? ”

146. See note 37 above.