Contemporary civil wars frequently involve numerous militant actors.1 Vying for scarce resources, these armed groups compete with rival organizations to maximize their share of support from local populations or external patrons. A prominent strand of scholarship posits that, in order to secure support, armed groups will attempt to outbid rivals to demonstrate organizational effectiveness and resolve. Scholarship on outbidding is “one of the most widely cited theories of terrorist groups’ motivations and actions.”2 According to this theory, competition may lead groups to increase the scale and scope of their violence as they seek to attain a “market share” of support.3 In demonstrating their resolve, the most violent groups are expected to attain scarce resources at the expense of their less violent competitors.
While existing research often focuses on violent escalation as a primary consequence of armed group competition, examples from disparate conflicts suggest that militant organizations may not always compete with rivals through increased levels of violence. In Peru, for example, the Túpac Amaru Revolutionary Movement sought to distinguish itself from the Shining Path by launching less violent, but highly symbolic, attacks.4 In another case, Jabhat al-Nusra jettisoned its previous use of indiscriminate attacks in Syrian urban centers to avoid alienating local support. In doing so, the group “differentiated itself starkly from its former ISIS umbrella, whose operations in Iraq and now in Syria frequently target civilians en masse.”5 During the insurgency against British rule in Mandatory Palestine, Haganah often avoided the use of terrorist attacks, unlike Irgun and Lehi, framing its more restrained acts of resistance as the best path to achieving an independent Jewish state.6 As exemplified by these cases, even when faced with a competitive environment, armed groups may have pressing incentives to adopt strategies besides violent outbidding to attain popular support.
This paper explores an alternative militant strategy to violent outbidding: strategic restraint. Strategic restraint occurs when an armed group differentiates itself from rivals by choosing to limit, rather than escalate, its use of violence. Armed groups pursuing strategic restraint leverage variations in the population’s preferences toward acceptable levels of violence. In doing so, armed groups can carve out a niche of support that differs from that of their rivals as well as avoid the negative backlash that may result from violent escalation. One important way armed groups may advertise their relative restraint is by restricting violence against local populations. More specifically, organizations can avoid the use of indiscriminate tactics and mass casualty attacks against civilians. Highlighting their focus on avoiding civilian casualties, armed groups may utilize organizational propaganda to further emphasize their restrained use of violence, condemning and distancing themselves from highly lethal attacks against civilians.
We argue that armed groups choose strategic restraint to differentiate themselves from other violent armed groups when there is a wide range of opinions among potential supporters about what is an acceptable level of violence against civilians. This variety of preferences creates space for an armed group to distinguish itself. When faced with a narrow range of preferences, armed groups are restricted to competing through the quantity rather than the characteristics of violent attacks, characteristics such as the level of violence against civilians. Shocks, such as state repression or popular backlash against insurgent violence, create opportunities for using strategic restraint by shifting the distribution of supporters’ preferences toward favoring or condemning violence against civilians. Following such a shock, a group will often demonstrate strategic restraint under two possible circumstances: either room opens in the “marketplace” and allows for the formation of a new restrained group seeking to differentiate itself from a violent incumbent, or a restrained group is forced to emphasize its restraint when faced with a new competitor that is escalating its level of violence.7 The first situation occurs when a shock shifts the distribution of preferences toward favoring more restrained violence, while the latter takes place when the distribution of preferences shifts toward more extreme forms of violence.
We examine these arguments by analyzing competition between jihadist groups in Algeria (1998–2004) and Yemen (2015–2021).8 Specifically, we analyze the behavior of the Salafist Group for Preaching and Combat (GSPC) in Algeria and al-Qaeda in the Arabian Peninsula (AQAP) in Yemen as they competed with rival jihadist groups for support. Conventional wisdom suggests that jihadist groups are the most likely ideological organizations to compete with rivals through escalatory violence.9 However, drawing on event-level data and Arabic-language jihadist publications, we highlight how the outbidding theory struggles to explain the behavior of the GSPC and AQAP. While the GSPC used vacant space in the Algerian marketplace of jihadist groups to differentiate itself through restraint, AQAP further emphasized its restrained violence when faced with a new violent competitor. Although competitors often employed brutality and launched large-scale attacks against local populations, the GSPC and AQAP strategically limited their violence against civilians to demonstrate their relative moderation. The GSPC and AQAP further highlighted their commitment to restraint by rhetorically condemning and distancing themselves from violence against civilians.
In making these arguments, this article contributes to several areas of research. First, it provides insight into how armed groups strategically respond to competition with rival militant groups. Although recent advances in the civil war and terrorism literatures explain armed group infighting or alliance formation, other strategic options available to organizations besides violent outbidding remain relatively understudied. Further theory-building efforts are needed to explain why armed groups may compete with rivals without engaging in violent escalation. Providing a greater understanding of the full menu of strategic options available to armed groups to attain support from local populations is critical to both scholars and policymakers. After all, civil conflicts across the world have grown increasingly fragmented in recent years, with more armed groups reportedly forming in the last decade than in the previous century.10
Although scholars of the outbidding theory predict that the existence of violent competitors will tempt even restrained groups to use more escalatory violence, we argue instead that restrained groups often possess ways of competing through advertising their restraint.
Contributing to these theory-building efforts, this paper builds on the growing body of political violence literature surrounding armed group restraint.11 To explain why some groups are able to exercise and may pursue restraint, previous research on restraint has generally focused on internal group-level attributes, such as initial resource endowments, the size of an armed group’s support base, investment in political education, and the relative power of a group within a movement.12 Others have emphasized aspects of the environment in which an armed group operates, including whether peace talks are in process,13 the effects of external shocks,14 or the total number of groups operating in the area.15 Broadening this literature, we illustrate how strategic interaction between armed groups may serve as a critical contributing factor to enabling restrained armed groups to enter and survive in competitive environments. Although scholars of the outbidding theory predict that the existence of violent competitors will tempt even restrained groups to use more escalatory violence, we argue instead that restrained groups often possess ways of competing through advertising their restraint.
Finally, this article’s focus on jihadist organizations provides insight into a prominent subset of armed groups operating in the international system. As Thomas Hegghammer states, “No other ideological family has fostered a set of militant groups as large, as mobile, and as resilient as the jihadist movement.”16 Rather than operate in isolation, multiple jihadist groups have competed for support in civil conflicts in Afghanistan, Algeria, Egypt, Indonesia, Iraq, Libya, Malaysia, Nigeria, Pakistan, Palestine, the Philippines, Somalia, Syria, and Yemen.17 Some of these groups have competing allegiances to al-Qaeda or the Islamic State (ISIL). While al-Qaeda and Islamic State “central” have suffered repeated setbacks due to increased counterterrorism pressure and loss of territorial safe havens,18 these transnational networks expanded by establishing affiliate groups in various geographical areas.19 Understanding the behavior of al-Qaeda and ISIL affiliates continues to be a pressing priority for U.S. policymakers. Reflecting the continued need to combat jihadist groups, the U.S. Department of State declared in March 2023 that “the United States remains firmly committed to working through the Global Coalition to Defeat Daesh/ISIS and its partners to ensure the terrorist group’s enduring defeat in the Middle East and its affiliates operating in Africa, Central Asia, and anywhere it seeks a foothold.”20
This paper proceeds as follows. First, we review the literature on relations between armed groups and outbidding. Next, we discuss why armed groups may utilize strategic restraint as a means of differentiating themselves from rivals. We then present our research design and case studies, underscoring how differentiation better explains the GSPC’s and AQAP’s restrained behavior — cases in which much of the literature would expect outbidding to occur. Finally, we summarize our findings and discuss various policy implications.
Armed Group Competition: Outbidding and Strategic Restraint
Rather than focusing solely on the interactions between militant organizations and the state, a growing body of literature explores relations between armed groups. Such research explores why armed groups with a common goal choose to fight each other or decide that their interests are better served through cooperation.21 Some organizations originally allied under the same common cause may fragment,22 while other groups previously at war with each other may establish an alliance to combat a common threat, underscoring the fluidity of relations between militant actors.23 Importantly, between the two extremes of a formal alliance and all-out war, there is a wide range of cooperative and competitive interactions between armed groups.
Armed groups in the same conflict commonly rely on a shared pool of resources and recruits. In crowded conflict environments, armed groups must identify ways to distinguish themselves from competitors to ensure that they secure adequate backing to survive.24 Facing competition from rivals, armed groups operating in fragmented environments may have incentives to alter their behavior. Seeking to win a greater portion of the “market share” of support, militants may outbid rivals by increasing the scale and scope of their militant activities.25 On one hand, increasing their levels of violence can signal organizational resolve and efficacy to local populations. On the other hand, failing to escalate in competitive settings can be costly for an armed group, as popular support may shift toward a more violent rival. When a group is faced with a competitor, Justin Conrad and William Spaniel argue that “producing little violence is inadvisable. Slacking off in this manner allows the other group to siphon off support by committing a few attacks and increasing its brand recognition.”26
Violent outbidding thus constitutes a prominent theory for explaining the consequences of armed group rivalry.27 Focusing on suicide bombings, for example, Mia Bloom notes that “if multiple insurgent groups are competing for public support, bombings will intensify in both scope and number as they become both the litmus test of militancy and the way to mobilize greater numbers of people within their community.”28 Scholars have used outbidding to help explain armed groups’ use of female suicide bombers and the targeting of children, civilians, and nongovernmental organizations.29 Moreover, competition may lead armed groups to increase their brutality, as portrayals of “gruesome violence transmit an unmistakable signal of the group’s commitment to its cause.”30 Recent research has also employed the concept of “intra-extremist outbidding” to explain why some right-wing actors openly praise jihadist beliefs and violence. Operating against rivals, right-wing extremists may “compete in endorsing the most extreme image of jihadist violence, regarding jihadists as successful trendsetters infamous for their brutality and shocking violence, that would attract the most devoted recruits.”31
To highlight their commitment to using increased levels of violence, organizations may also alter their propaganda while competing with rivals. Armed groups operating in competitive environments may have greater incentives to formally claim acts of violence. After all, potential supporters can struggle to identify the perpetrators of different acts of violence in conflict settings populated by multiple militant organizations. Claiming credit for attacks may solve this issue, “enabling those that commit acts of terrorism to reap the organizational benefits of violence.”32 Armed groups may use propaganda to claim responsibility for violence in their efforts to signal “that the group is strong and worthy of support.”33
The incentives to take credit for acts of violence may also lead groups to claim attacks that they did not commit. Indeed,
when there are multiple competing groups and an attack goes unclaimed, there may be additional incentive to free ride and not attack but to claim credit for another’s violence. Taking credit for another group’s work can cause doubt among the population over the effectiveness of the rival.34
Such incentives to claim attacks might be present even when operations result in high levels of civilian casualties.35 Competition between armed organizations for scarce resources can thus lead groups to engage in violent outbidding and employ propaganda to further underscore their commitment to using high levels of violence.
Violent escalation may also result in attacks that transgress cultural norms or incite harsh government retaliation.
Despite outbidding’s continued prominence in political violence scholarship, however, recent studies employing qualitative and quantitative methods have found mixed empirical support for its theoretical propositions.36 These studies suggest that armed groups may have powerful reasons to forgo violent escalation. For instance, previous scholarship has shown that groups that use terrorism against civilians are less likely to achieve their long-term goals.37 Moreover, the manner in which groups employ violence is key for internal cohesion. Armed groups that use certain targeting strategies or indiscriminate tactics may alienate organizational factions and cause splintering. For example, Boko Haram’s suicide attacks against Muslim civilians may have influenced Ansaru’s split from the group in 2012: The latter distributed flyers declaring itself a more “humane” alternative to Boko Haram.38
Violent escalation may also result in attacks that transgress cultural norms or incite harsh government retaliation.39 By triggering a backlash, violent attacks may lead to a shrinking pool of supporters who favor high levels of violence or push civilians to provide information to state forces.40 For example, the 1997 Luxor massacre, in which 62 foreign tourists and Egyptians were killed, generated significant condemnation against Egyptian Islamist groups and increased support for the regime’s counterterrorism operations.41 In Uruguay, the Tupamaros gained recruits by carrying out more violent attacks, but it lost much of its broader popular support.42
While outbidding sheds critical light on the competitive dynamics between armed groups in some conflict settings, armed groups also have strategic incentives to restrain their violence. In doing so, armed groups can differentiate themselves from rival organizations, forgoing the opportunity to engage in violent outbidding. Rather than losing critical popular backing, armed groups that utilize strategic restraint may carve out a niche among segments of a support base that prefer lower levels of violence against civilians.
Strategic restraint is a strategy of competition that armed groups use to differentiate themselves from rivals by limiting, rather than escalating, violence. Armed groups engage in restraint by strategically applying the “internal brakes” on their use of violence, imposing restrictions on specific acts of violence.43 To be considered restrained, armed groups must have the capacity to produce more violence than they choose to employ. They must intentionally curb their use of violence rather than have that choice forced upon them by external factors.44
Our definition of strategic restraint emphasizes the role of differentiation strategy. A common strategy in the literature on industrial organization, differentiation is a form of soft competition where each organization carves out a portion of the support base for themselves by offering different goods than their rivals.45 When it comes to competing armed groups, the preferences of potential supporters exist on a spectrum, ranging from those who favor attacks resulting in few civilian deaths to those who prefer high-lethality attacks, regardless of the target. Although measuring the preferences of civilian supporters is difficult, various examples indicate a variety of preferences for the level of violence used by armed groups.46 This allows an armed group to differentiate its brand from rival groups.47 An armed group may differentiate its brand through its governance, taxation, or ideology.48 However, given the drawbacks associated with violent outbidding and the population’s varying preferences, armed groups may have an incentive to strategically limit certain elements of their violence to differentiate themselves from competitors. One way an armed group may do so is by curbing violence against civilians.49
Although groups may also signal restraint by forgoing certain tactics, such as suicide bombings, we focus on armed groups’ levels of violence against civilians for multiple reasons. First, while potential supporters may favor violent resistance against the state, they may have disparate preferences about the acceptable levels of violence against civilians. For instance, portions of the Provisional Irish Republican Army’s support base in Northern Ireland demanded retaliatory attacks against civilians. However, civilian casualties ultimately eroded the group’s support, with an organizational spokesperson even admitting in early 1989 that “many civilians died in operations which dented the confidence of some of our supporters.”50 Indeed, potential supporters care not only about the quantity of attacks an armed group conducts, but also about the quality of militant operations.51 For example, recent findings highlight how Islamic State propaganda showcasing brutal violence decreased general endorsement for the organization, with such content only gaining support among the group’s most extreme supporters.52
While highly lethal attacks against civilians — like suicide bombings in crowded spaces — can underscore an organization’s efforts to outbid during competition, armed groups pursuing restraint may lower their levels of violence against civilians to differentiate themselves from such groups. This is not to say that armed groups pursuing restraint will never target civilians. Rather, organizations attempting to restrain their use of violence may avoid attacks that cause mass civilian casualties and the use of indiscriminate tactics in highly populated areas. Organizations that are seeking to limit their use of violence against civilians may focus instead on major operations against state forces. By attacking regime and military targets, an armed group can still effectively signal its resolve to achieve organizational goals while avoiding the negative backlash associated with targeting civilians in large-scale attacks. For instance, the leftist Farabundo Marti National Liberation Front in El Salvador often launched attacks against economic and infrastructure targets, strategically avoiding areas that might generate high civilian casualties and, consequently, lower support for the group.53
For strategic restraint to be effective, however, an armed group must also communicate this restraint to potential supporters, voicing its commitment to avoiding large-scale attacks that cause significant civilian casualties. In addition to using violence, rhetorically differentiating their ideology, tactics, and demands provides another avenue through which armed groups may set themselves apart from rivals.54
Armed groups pursuing this strategy may deny having links to certain attacks or even publicly apologize for excessive violence against civilians due to an accident or a reckless subordinate.
Armed groups pursuing strategic restraint possess at least three ways to distinguish themselves rhetorically from less restrained rivals. First, armed groups may emphasize their restraint by acknowledging the lack of harm that their attacks cause to civilians. Emphasizing restraint often involves clearly communicating whom or what the group views as legitimate targets and whom or what they consider to be off limits. Second, in contrast to outbidding’s race to claim credit for violent attacks, armed groups pursuing strategic restraint will explicitly distance themselves from highly lethal attacks against civilians. Armed groups pursuing this strategy may deny having links to certain attacks or even publicly apologize for excessive violence against civilians due to an accident or a reckless subordinate. Denying responsibility or publicly apologizing for violence would be unexpected under a violent outbidding strategy, which emphasizes an organization’s ability to conduct highly lethal attacks. Finally, armed groups pursuing strategic restraint might also publicly condemn a rival’s use of large-scale attacks and indiscriminate tactics against civilians. Public condemnation allows restrained groups to appropriate a rival’s violent attacks to remind potential supporters that an alternative group exists and may be more worthy of support. The MRTA, for instance, operated a radio station, Radio 4 de Noviembre, and a publication, Venceremos, which often criticized the more violent Shining Path for its use of violence.55
Overall, armed groups may possess powerful incentives to differentiate themselves from competitors by pursuing strategic restraint. Rather than remain silent, armed groups can signal their commitment to restraint through propaganda. In addition to analyzing the ways in which groups may advertise their use of restraint, it is critical to understand the conditions under which strategic restraint may occur.
When Does Strategic Restraint Occur?
There must be at least some supporters who prefer low levels of violence against civilians for strategic restraint to be a viable option. As the range of preferences regarding violence expands, the more likely it is that an armed group will be able to differentiate itself through strategic restraint. If the range of supporter preferences is narrow, then an armed group will have difficulty distinguishing its own level of violence from that of its rivals. When faced with a narrow range of preferences, the quantity of attacks committed by each armed group becomes more important for winning popular support. For instance, in the prominent case of outbidding in Palestine, most of the armed groups competed for only a narrow fraction of Palestinian public opinion that favored greater violence. Given the narrow range of supporters’ preferences, these groups could only distinguish themselves by the quantity of suicide attacks launched.56
Identifying when a restrained group can compete through advertising its restraint is critical as theories of outbidding note that even previously restrained armed groups may be tempted to use more escalatory violence within a competitive environment.57 Differentiation through strategic restraint will commonly occur when a new armed group seeks to differentiate itself from a violent incumbent, or when a restrained group faces a new competitor that uses more violence against civilians than existing armed groups. Until a competitor comes along, armed groups will pursue a violent or restrained strategy depending on the specific distribution of supporter preferences.58 However, when shocks occur that expand the pool of supporters or shift the distribution of preferences, armed groups will be forced to compete with new rivals by either choosing to escalate or restrain their use of violence. In other words, rather than always leading to escalatory violence, the arrival of more armed groups may cause existing groups to pursue strategic restraint when competing for support from a population with a wide range of preferences. Additionally, identifying what factors may shift supporters’ preferences can help to predict when armed groups will gain or maintain a niche of supporters by advertising their restraint.
Shocks or triggers have been found to cause rises in violence by altering norms or removing moderates from leadership positions.59 However, such shocks may also create incentives for armed groups to engage in strategic restraint by creating a vacant space for armed groups that favor relatively moderate levels of violence. Although it is beyond the scope of the paper to discuss every possible factor that may shift the distribution of preferences, two stand out: escalatory violence by an incumbent armed group and the state’s willingness to negotiate or use repression. Strategies pursued by existing armed actors in a conflict may shift potential supporters’ preferences toward favoring lower or higher levels of violence. As noted above, the use of highly violent attacks by an incumbent armed group may generate backlash over time that opens space for the entry of a new armed group that will pursue strategic restraint. A shift in the preferences of the support base toward moderation will allow armed groups possessing ideologies or characteristics that favor more moderation in their use of violence to enter and compete with existing armed groups.60 Alternatively, a collapse of negotiations or renewed government repression may radicalize the population,61 causing them to accept higher levels of violence. This may make room for a violent competitor, forcing a restrained incumbent to emphasize its limited use of violence against civilians to maintain its niche of supporters.
Although an incumbent armed group may be tempted to simply follow shifts in the population’s preferences, it will struggle to quickly switch from a strategy of escalating violence to strategic restraint. Any armed group considering a strategic shift would need to reorient its recruitment and training practices to develop the means to control its use of violence and reject recruits who favor heightened levels of violence against civilians.62 Unsanctioned attacks by subordinates will undermine a group’s ability to compete through strategic restraint. If supporters cannot distinguish the levels of violence used by different groups, none of the groups will be able to differentiate themselves given their attacks do not differ from attacks conducted by rival groups. Additionally, such organizational changes may be unpopular with many current members who joined the group due to their use of high levels of violence or an ideology that emphasizes violence.63 Many of these arguments similarly apply to a group that previously engaged in strategic restraint and is now attempting to increase its level of violence. Therefore, given incumbent armed groups cannot quickly switch their strategy to respond to the change in preferences, they will struggle to prevent the entry of new armed groups seeking to fill the opened space.
Armed groups choose strategic restraint to differentiate themselves from other violent armed groups when there is a wide range of preferences among potential supporters with regard to violence against civilians.
It is important to acknowledge that the theory presented here seeks to explain how and when restrained armed groups may effectively compete for support with more violent groups rather than why particular groups favor restraint. A specific armed group may prefer restraint due to various factors, such as its leadership, ideological position, or past experience. However, whether a new restrained group enters the scene due to ideological beliefs or because its leadership views excessive violence against civilians as ineffective is outside the scope of this paper. Rather, the restraint theory seeks to identify when armed groups will have sufficient “market share” to begin competing for support through restraint and why an existing restrained group is able to maintain a niche of supporters while forgoing violent outbidding with newly formed rivals. In a competitive environment where the distribution of potential supporters largely favors high levels of violence, an armed group will struggle to survive or will face pressure to escalate, even if it possesses characteristics predisposing it to employ restraint.
In summary, armed groups have an incentive to use strategic restraint to differentiate themselves from rival groups that engage in violent outbidding. Even when faced with a rival group that is increasing its levels of violence, an armed group that is using restraint would be expected to conduct few high-lethality attacks against civilians despite possessing the capabilities to do otherwise. A group pursuing restraint may brand itself as qualitatively different from its more violent rivals. Moreover, a group that is using strategic restraint will rhetorically emphasize its commitment to limited violence, distance itself from large-scale attacks against civilians, and condemn attacks by rivals using such violence. Armed groups choose strategic restraint to differentiate themselves from other violent armed groups when there is a wide range of preferences among potential supporters with regard to violence against civilians. In addition, shocks that shift the distribution of supporters’ preferences provide space within which new groups can differentiate themselves from an incumbent armed group. Rather than lose support, groups utilizing strategic restraint may carve out a niche of backing among supporters who prefer lower levels of violence. We illustrate these theoretical points below by exploring competition between jihadist organizations in Algeria and Yemen.
In this paper, we examine two prominent examples of armed group competition where existing research predicts that militant organizations will outbid their rivals through escalatory violence. We analyze competition between jihadist rivals, exploring the behavior of the Salafist Group for Preaching and Combat (GSPC) as it competed with the Armed Islamic Group (GIA) in Algeria (1998–2004), as well as al-Qaeda in the Arabian Peninsula’s (AQAP) rivalry with the Islamic State in Yemen (ISY) (2015–2021). We focus on the behavior of jihadist groups within Algeria and Yemen for three reasons. First, current research considers jihadist groups as “most likely” to outbid. Second, it was necessary to focus on groups with shared ideologies operating across similar geographic areas. Finally, existing political science scholarship often overlooks conflict dynamics in Algeria and Yemen. We explore each of these points in greater depth below.
First, conventional wisdom in the terrorism literature holds that jihadist organizations are among the most violent armed groups. For example, James Piazza finds that religious groups have a much higher average number of victims per attack than leftist, rightist, and nationalist groups. Jihadist groups linked to al-Qaeda are key in driving these findings, as these organizations are more likely to engage in highly lethal operations and cause a disproportionate number of casualties compared to other Islamist and non-religious organizations.64 Jennifer Carson and Matthew Suppenbach further explore the relationship between ideology and lethality, stating that the global jihadist movement has “become a dominant manifestation of terrorism” in recent years with “a distinct ability to kill and maim civilians.”65 More recently, Ido Levy finds that jihadist groups are more lethal and efficient killers than non-jihadist organizations.66
Scholars explain these findings related to religious groups and violence in various ways. For one, while secular groups may strive to “win the hearts and minds” of local populations, there is an assumption that religious groups may “not crave popular approval for their acts because they expect instead to obtain spiritual reward, making them even less inhibited when it comes to committing acts likely to yield high casualty rates.”67 Second, religious groups have highlighted their commitment to indiscriminate and lethal tactics. Indeed, numerous jihadist groups have used tactics like suicide bombings under the belief “that suicide operations against ‘infidels’ and ‘apostates’ represent the ultimate form of devotion to God and the optimal way to wage jihad.”68
With a commitment to conducting lethal attacks and a belief in a higher power, religious organizations may be especially prone to attempt to outbid rivals through escalatory violence. Indeed, Stephen Nemeth finds that nationalist and religious groups often respond to competition with increased levels of terrorism, suggesting that the latter type of organizations “are not concerned with public judgment but that of the hereafter.”69 Additionally, recent scholarship notes that competition among religious factions often leads to violent outbidding, as religious groups seek to enhance their credibility and visibility to augment ties to external networks of support.70 Scholarship finds that jihadist organizations may also respond to competition with increased levels of violence. For instance, Jama’at Nusrat al-Islam wal-Muslimeen reportedly “intensified civilian targeting to communicate strength and outbid” the Islamic State in Greater Sahara.71 Moreover, Megan Farrell contends that jihadist organizations often attempt to raise the “shock value” of their operations as they attempt to attain an increased market share of transnational support during competition.72
This is not to say that all jihadist groups employ extreme forms of violence across space and time. Indeed, recent scholarship contends that jihadist groups “are an especially rule- and norm-bound class of insurgents, and their legitimacy depends on their adherence to their Islamist ideals.”73 However, we choose to focus on the behavior of armed groups that scholars and policymakers frequently consider to be the “least likely” to employ strategic restraint while competing with rivals because of their desire to gain attention from transnational networks, their belief in a higher power, and their frequent use of indiscriminate tactics like suicide bombings. Indeed, Hegghammer notes that the academic literature often perceives jihadist groups as “more extremist and intransigent than other groups.”74 If the argument for strategic restraint has merit among the least likely cases, it would indicate a greater generalizability and would enhance our understanding of the competitive dynamics among armed groups in fragmented conflict settings.
Examining civil wars in Algeria and Yemen thus allows us to analyze outbidding and strategic restraint in relatively understudied conflicts in the Middle East and North Africa.
The case studies presented below include two sets of rival jihadist groups that were active in overlapping geographic areas and that operated within the same “ideological market.”75 Recent critiques of outbidding note the frequent creation of an “artificial market of competition” in single-country studies of armed group competition. By pooling all groups in a state together, single-country studies often give the false impression that the groups are all in competition with one another. For example, such an approach can indicate “that separatist groups operating in northeast India are in competition for the same resources as Indian Islamic groups operating near Kashmir.”76 To avoid this methodological misstep, we compared the behavior of groups that were active in close geographic proximity. Additionally, we compared groups that possessed comparable ideologies to ensure that the groups were competing for support among similar populations.
In this paper, ideology refers to a systematic set of ideas that identifies a constituent group, voices grievances facing this group, articulates objectives for this group, and delineates a program of action.77 Importantly, key differences may exist among groups that possess a jihadist ideology. For example, jihadist organizations may have different enemy hierarchies and varying territorial aspirations. They may also frame their struggle in disparate ways.78 Despite inherent variation in the views of various organizations, we analyzed jihadist groups with a significant degree of ideological overlap. These groups espoused similar grievances and territorial aspirations, adopted a program of violence to achieve organizational goals, and professed a Salafist identity. For instance, the GIA and GSPC were both “state oriented” groups that denounced the tyrannical governance of “apostate” local rulers and sought to change the social and political organization of the Algerian state through socio-revolutionary activism.79 In contrast, al-Qaeda and ISIL have rejected the international order and adopted pan-Islamist identities while seeking to create political entities that transcend current international borders.
Finally, the conflicts in Algeria and Yemen are understudied in political science research, making them ideal candidates. An examination of leading comparative politics and international relations journals underscores that political violence research on the Middle East and North Africa primarily focuses on countries like Iraq, Israel, Palestine, and Syria.80 Moreover, Brian Phillips and Kevin Greene reviewed five prominent conflict journals between 1990 and 2015 and found that, despite experiencing substantial conflicts during this time, Algeria and Yemen were rarely mentioned in article abstracts or were the subject of case studies in conflict research.81 Noting that the articles in the publications focus on conflicts in a relatively small number of states, they ask: “[A]re we over-learning the lessons of Northern Ireland (for example), and ignoring the lessons of Ethiopia or Algeria?” 82 Examining civil wars in Algeria and Yemen thus allows us to analyze outbidding and strategic restraint in relatively understudied conflicts in the Middle East and North Africa.
We examine these case studies through the use of process tracing. An important tool for conducting qualitative research, process tracing involves analyzing “processes, sequences, and conjunctures of events within a case” by utilizing assorted pieces of evidence.83 Recent research has underscored the benefits of using process tracing to evaluate competing theories surrounding armed group behavior and tactics.84 Accounting for variation in armed groups’ employment of violence over time, process tracing allows for the comparison of strategic restraint with the primary alternative theory of violent outbidding. As a critical element of process tracing, we draw on multiple pieces of evidence to make our arguments. First, we utilize data from the Armed Conflict Location and Event Data Project (ACLED). Data from this project provides an overview of the temporal variation in attacks conducted by the four groups examined below.85 While quantitative analyses of outbidding often rely solely on event data, this information only paints a partial picture of armed group behavior.
Seeking to understand more fully why jihadist groups may strategically employ restraint, we examine numerous Arabic primary sources, including organizational statements, newspapers, and videos, to explore how the GSPC and AQAP responded to competition in their propaganda. Organizational publications are key for supporting the theory of strategic restraint. Not only do these sources offer unique insight into groups’ conceptions of their ideology, goals, and strategy, but they also shed light on why militant groups employ certain forms of violence and attack specific targets.86 Without examining the strategies outlined by groups in their own words, it would be difficult to know, based on event data alone, when and why militant organizations seek to strategically restrain their violence or to violently outbid. After all, the capacity of armed groups may not mirror their willingness to launch certain attacks, limiting their presence in conflict databases.
It is important to note that this paper’s methods and data have different limitations. For one, propaganda produced by the GSPC and AQAP may attempt to downplay mistakes, glossing over missteps to focus on military triumphs. Moreover, we collected a number of GSPC and AQAP publications via jihadology.net, a well-known website that researchers may use to safely access a corpus of jihadist primary sources.87 Given that this website does not contain every piece of propaganda published by these organizations, relying on a limited sample of publications could potentially bias our findings. Additionally, despite examining numerous Arabic primary sources, this paper does not engage with the wealth of French-language media and research that focuses on conflict dynamics in Algeria and the Sahel.88 Finally, utilizing case studies limits our ability to generalize our findings beyond the contexts of Algeria and Yemen.
Despite these limitations, however, our research design furthers our article’s goal of theory-building. It does so by analyzing understudied conflicts in Algeria and Yemen, selecting “least likely” cases for restraint to occur, and examining Arabic-language primary sources. Engaging with various publications produced by these groups over time, as well as additional jihadist documents from various sources, provides crucial insight into organizational behavior. Furthermore, this paper illustrates the need for scholars and policymakers to account for the conditions under which competition may produce strategic restraint in future analyses of militant behavior.
Jihadist Competition in Algeria
Various jihadist groups formed over the course of the Algerian civil war (1992–2002), a conflict that resulted in an estimated 150,000 deaths. Among the most powerful organizations, the Islamic Salvation Front and its armed wing — the Islamic Salvation Army — competed with the Armed Islamic Group (GIA) for popular support and hegemony within the jihadist movement. In contrast to the Islamic Salvation Front, the GIA formed as an amalgamation of disparate Islamist groups in response to the escalation of violence after the Algerian military’s cancellation of elections in January 1992.
The GIA initially garnered support from prominent figures and groups in the global jihadist community, such as Osama bin Laden and the Libyan Islamic Fighting Group, to help achieve its goal of establishing an Islamic state in Algeria by means of “total war.”89 Buoyed by domestic and international support, the GIA rose to prominence because of its military successes in the early years of the Algerian conflict. In doing so, the GIA became “the world’s most successful jihadi group — the group that came closest to dominating one side of a civil war and to overthrowing” a local regime in the 1990s.90 Importantly, Alexander Thurston states that the GIA also became “the first armed group to fully subordinate jihadi ideology to Salafi theology” as it professed its theological purity.91 Exemplifying this point, the appearance of “jihadi-salafi” and related terms in jihadist texts and organizational names remained relatively rare prior to 2003.92 Following the GIA’s early momentum, however, the group experienced salient behavioral changes that played a central role in overturning its initial successes.
The GIA’s indiscriminate tactics, massacre of civilians, and broadening scope of legitimate targets of violence incited backlash at home and abroad.
The issue of civilian targeting became a key point of contention among Algerian jihadists during the country’s civil war. On one hand, the Islamic Salvation Army and the Islamic Salvation Front issued statements prohibiting attacks against civilians. “[R]ather than seek to outbid the GIA with escalatory violence, the AIS [Islamic Salvation Army] sought to differentiate itself by insisting on targeted violence,” focusing its operations on security forces and regime personnel.93 In contrast, the GIA increasingly expanded its targeting of different types of groups during the mid-1990s. The GIA’s attempts to control various forms of social behavior furthered its attacks against civilians “under the pretext of their ‘un-Islamic’ behavior, such as smoking, drinking alcohol or failing to perform their prayers: in the eyes of the GIA, such failure to abide by Islamic strictures deserved punishment by death.”94 Over time, the GIA deemed anyone who did not support it to be infidels and supporters of the Algerian regime.
Emphasizing its willingness to utilize extreme methods of violence against a wide range of civilian targets, the GIA began attacking foreign nationals, journalists, teachers, and students. Addressing foreign nationals residing in Algeria, for instance, the organization stated: “Leave the country. We are giving you one month. Anyone who exceeds that period will be responsible for his own sudden death. There will be no kidnappings and it will be more violent than in Egypt.”95 Additionally, the GIA deemed any participation in educational institutions as an expression of support for the regime, leading to assassinations of teachers and bomb and arson attacks against schools.96 Importantly, the GIA’s campaign of violence was not limited to Algeria. The organization launched multiple attacks against French targets, such as the 1994 hijacking of Air France Flight 8969, in an attempt to internationalize the conflict.97 Antar Zouabri — the GIA’s leader from 1996 until 2002 — began “waging total war against Algerian society in order to overwhelm and subdue it” as he broadened the scope of the GIA’s attacks through the use of takfir (the act of Muslims declaring other Muslims to be unbelievers).98 In doing so, Zouabri eventually condemned the entire Algerian population of being guilty of apostasy, making them legitimate targets of violence.99
The widening scope of legitimate targets culminated in large-scale massacres against Algerian civilians, especially over the course of 1997. Claiming credit for one such massacre, a GIA communiqué claimed that the organization “follows the traces of those apostates in the cities, villages and deserts and then wipes them out and destroys their fields.”100 Rather than using bombs or firearms, massacres were often carried out through face-to-face violence. Assailants “wielded knives, machetes, and swords, necessitating close proximity to the victims. Ordinary citizens were maimed, decapitated, and burned alive at an alarming rate.”101 By employing overt brutality against civilians, the insurgents signaled “that although death at their hands might be less certain than death at the hands of the army, it will definitely be more brutal: more painful (through the use of knives and axes), more comprehensive (including entire families), transgressive of taboos (mutilation of dead bodies), etc.”102
The GIA’s indiscriminate tactics, massacre of civilians, and broadening scope of legitimate targets of violence incited backlash at home and abroad. The use of such wide-scale violence led many to call into question the organization’s legitimacy,103 as the popular mood in Algeria shifted “dramatically against the Islamic movement in general and the regime capitalized by portraying all Islamists involved in the conflict as part of a single, ultra-violent jihad movement.”104 Indeed, the Algerian regime recognized the detrimental impacts of the GIA’s actions: “By going easy on the GIA, the military hoped to induce the AIS [Islamic Salvation Army] to surrender by causing it to lose its base of support as a result of the impact of GIA violence.”105
Besides the Algerian populace, the GIA’s extreme use of violence alienated numerous group members and actors in the international jihadist community. According to one former Salafist Group for Preaching and Combat (GSPC) member, bin Laden had encouraged fighters to defect from the GIA in an effort to improve the jihadist movement’s image following the GIA’s campaign of massacres against Algerian civilians.106 Moreover, Abu Musab al-Suri, a leading jihadist theoretician and military strategist, retracted his support from the GIA in 1996, as “the high expectations which al-Suri and his fellow jihadis had attached to the Algerian jihad” had been “thoroughly quashed by the spiraling barbarism of the GIA’s violence” conducted under the leadership of Djamel Zitouni and Zouabri.107 Rather than always agreeing on appropriate levels of violence, domestic and international criticism of the GIA’s tactics reflected intra-Islamist divides over issues like the use of takfir, targeting Muslims in militant operations, and appropriate tactics.
The uptick in the GIA’s massacres in 1997 coincided with key developments within the international jihadist movement in the late 1990s and early 2000s. Domestically, the Islamic Salvation Army altered the course of the Algerian conflict by declaring a ceasefire in October 1997, formally disbanding in 2000. In contrast to the GIA, the Islamic Salvation Army viewed “indiscriminate violence as a threat to its political project and declared a unilateral ceasefire to avoid a two-front war with the regime and the GIA. … [O]nce it became clear that armed struggle was damaging Islamists’ standing with the public, it made little sense to continue with the insurgency.”108 Outside of Algeria, other prominent jihadist organizations also agreed to lay down their weapons in 1997. For instance, the Islamic Renaissance Party also called for a ceasefire and resolution to the conflict in Tajikistan. Moreover, in a shocking move, leaders of the Islamic Group — Egypt’s largest jihadist group — declared a unilateral ceasefire in July 1997. In addition to calling for a cessation of military operations, Islamic Group leaders published numerous books to ideologically legitimize the organization’s deradicalization process.109
Ongoing debates in the jihadist community surrounding demilitarization, targeting strategies, and the appropriate use of violence occurred as the GIA continued to see numerous defections over the second half of the 1990s.
In contrast to these demilitarization efforts, bin Laden announced his “World Islamic Front Against the Jews and Crusaders,” in February 1998. Challenging the jihadist movement’s traditional focus on overthrowing local governments, bin Laden called for prioritizing attacks against the “far enemy” over operations targeting apostate regimes in Muslim majority countries (the “near enemy”).110 However, not all jihadists supported unrestrained violence against the United States. Even prominent al-Qaeda leaders, such as Saif al-Adel and Abu Hafs al-Mauritani, reportedly opposed the 9/11 attacks.111 Following 9/11 and the 2003 U.S. invasion of Iraq, Abu Musab al-Zarqawi’s tactics in Iraq received harsh criticism from figures within the jihadist movement, such as Ayman al-Zawahiri, Abu Muhammad al-Maqdisi, and Abu Basir al-Tartusi.112 Years later, Sayyid Imam al-Sharif, considered by some to be “the most influential ideologue in Jihadism after Sayyid Qutb,”113 provoked significant debate within the jihadist community in 2007 when he published the Document for Guiding Jihad in Egypt and the World.114 Recanting many of his previous views, al-Sharif condemned the 9/11 attacks, proscribed attacking foreign tourists in Muslim majority countries, and established strict limitations on declaring takfir.115 Ongoing debates in the jihadist community surrounding demilitarization, targeting strategies, and the appropriate use of violence occurred as the GIA continued to see numerous defections over the second half of the 1990s.
Within this milieu, prominent GIA subcommanders, such as Hassan Hattab, Nabil Sahrawi, and Mokhtar Belmokhtar, splintered in what Thurston terms a “field commanders’ revolt” against the GIA’s central leadership.116 Critically, the withdrawal of the more restrained Islamic Salvation Army had opened the door for a jihadist organization pursuing strategic restraint to enter the marketplace of Algerian armed groups. Through its founding, which is commonly dated to 1998, the GSPC helped to fill “a vacant space” in the conflict’s landscape due to “the radicalism of the GIA on the one hand and the truce declared by the AIS [Islamic Salvation Army] in October 1997.”117
Enter the Salafist Group for Preaching and Combat
Operating in the same ideological market as the GIA, the GSPC adopted similar political and religious positions to the GIA as it formed in the late 1990s. For one, the GSPC declared an unwavering commitment to achieving its long-term objectives of reestablishing the caliphate (al-Khilafa al-Rashida).118 To accomplish this goal, the organization sought to topple the Algerian regime, adopting the slogan that there could be “no dialogue, no truce, no reconciliation” (la hiwar, la hudna, la musalaha) with Algeria’s tyrants.119 Moreover, the GSPC maintained the GIA’s Salafist ideals. In its founding charter (al-Mithaq), the GSPC proclaimed its commitment to Salafism. Expressing a desire to return to the “correct” Salafist method (manhaj),120 the GSPC stated that its creed was that of “the pious predecessors” (al-Salaf al-Salih), or the first three generations of Muslims that are regarded as exemplifying the correct way to live and whose behavior Salafists strive to closely emulate.121 Despite striving to topple the Algerian regime, promoting the formation of an Islamic state, and adopting a Salafist identity, the GSPC differentiated itself from its predecessor by vowing to correct the GIA’s strategic mistakes.
Attempting to regain support and move beyond the GIA’s past behavior, the newly formed GSPC denounced the GIA’s uncontrolled violence. While clearly describing the worthiness of its cause, the GSPC proclaimed that “the end does not justify the means” (al-Ghaya la tubarriru al-Wasila). On this point, the GSPC vowed to avoid civilian casualties while concentrating its military operations on the Algerian regime, seeking to return to the original goals of the Algerian jihad prior to the GIA’s descent into depravity. For instance, the GSPC posted an early communiqué in the eastern district of Algiers, a city in which the GIA traditionally had a stronghold,122 condemning the crimes of the GIA, “which is still shedding the blood of innocent people in massacres.”123 Sahrawi — who led the GSPC from 2003 to 2004 after participating in a coup against Hattab124 — continued to echo this sentiment in 2004, stating that the GSPC had formed as an extension of the GIA and its method before the latter’s “deviations” (al-Inhirafat wa al-Zaig).125 Avoiding the GIA’s missteps meant that the GSPC would not rashly charge Muslims with apostasy. The GSPC claimed that, by utilizing mass violence against civilians, the GIA had violated the honor of “chaste women, spilled forbidden blood, and looted property, unjustly and in a way that is approved neither by Shariah nor by logic.”126
The GSPC grew in size after its formation, quickly adding several thousand fighters to its ranks.127 The GSPC became particularly strong in the mountainous Kabylia region east of Algiers and in regions in the country’s south.128 Previously, the GIA had “dominated the strategic main roads” in Kabylia and other areas in central and eastern Algeria to bolster its war economy.129 Financially, the organization was able to generate funding through diverse revenue streams, including various forms of criminal activity in the Maghreb and Europe.130 Despite increasing its financial and military capabilities, the GSPC’s early media activity remained sporadic. “Compared to other jihadist-inspired groups,” Manuel Torres Soriano states that “the GSPC was among the last to realize the need for major communicative actions as part of its strategy.”131 However, the GSPC continued to rhetorically underscore its commitment to avoiding civilian causalities in its organizational propaganda after its founding.
Despite its relative lack of media experience, the GSPC emphasized its commitment to restraint in documents published following the release of its charter. For instance, the organization professed its innocence on multiple occasions with regard to bombings that targeted public spaces. In 2001, the GSPC denied links to an explosion at a bus station in Tafourah that killed one individual and injured others, reiterating the group’s “enduring covenant” (‘ahdiha al-Thabit) not to target the Algerian populace in its operations.132 A year later, Yahya Abu al-Haitham — emir of the GSPC’s Zone 2 — proclaimed the organization’s innocence after a bombing in a market in Bejaia, emphasizing that the group avoids harming the Algerian people.133 Rather than rush to claim different attacks, the GSPC distanced itself from acts of violence that transgressed its commitment to strategic restraint. This is the opposite of what the theory of outbidding would predict: that, when faced with competition from a rival group, the GSPC had incentives to escalate in order to signal resolve to local populations and to adopt more extreme positions in crowded conflict settings, despite having articulated a strategy of restraint.134
Rather than outbidding through escalation, the GSPC restrained its use of violence against civilians during its rivalry with the GIA, operationally distancing itself from its predecessor’s excesses and proclaiming its innocence from attacks that violated its commitment to relative moderation.
Between January 1997 and December 2004, the GIA conducted a total of 453 attacks while the GSPC launched 235.135 Comparatively, a much larger proportion of GIA’s attacks were against civilians (79 percent) compared with the GSPC (17 percent). The number of civilian fatalities also differed significantly. The GIA recorded 6,142 civilian fatalities across 356 attacks (17.3 fatalities per attack) compared to the GSPC’s 105 civilian casualties across 41 attacks (2.7 fatalities per attack). While the GIA’s more violent strategy resulted in a high number of civilian causalities, the data underscore the GSPC’s operational commitment to avoiding competing with the GIA through high-lethality attacks against civilians.
Table 1 provides an overview of the 10 deadliest attacks conducted against civilians between 1997 and 2004 by the GIA and the GSPC. Consistent with our argument of strategic restraint, all of the deadliest attacks against civilians were committed by the GIA.136 Importantly, the GSPC’s relatively low-scale violence against civilians was unlikely to have been a product of military inexperience, as many of the GSPC’s members had fought for years under the GIA’s banner prior to defecting. Moreover, the GSPC showcased its capacity to attack Algerian police and military personnel through a variety of tactics, such as roadside bombings and ambushes. Throughout the early 2000s, total fatalities caused by GSPC attacks dramatically increased, even as the proportion of civilians targeted remained relatively low. GIA continued its commitment to targeting civilians throughout its final years, when it was competing with the GSPC. Rather than being held back by its restrained use of violence, the GSPC grew in size after its formation and continues to outlive the GIA, which became inactive in 2004. While the extent of the GSPC’s support was uncertain, Luis Martinez states, “It certainly seems unlikely that groups which commit crimes against villages could enjoy the public’s protection. Yet the Salafists’ ability to expand shows that they are not excluded completely from the public sphere.”137
Table 1: Ten Deadliest Attacks Against Civilians (Algeria)
|GIA||Sidi M'Hamed Benaouda||103||Massacre|
In sum, the GSPC strategically focused its early violence against regime targets during its competition with the GIA, filling a vacancy after the shock of the Islamic Salvation Army’s ceasefire. The latter’s decision to lay down its arms opened the door for a new group to appeal to supporters of the fight against the Algerian regime through less extreme forms of violence. Rather than outbidding through escalation, the GSPC restrained its use of violence against civilians during its rivalry with the GIA, operationally distancing itself from its predecessor’s excesses and proclaiming its innocence from attacks that violated its commitment to relative moderation.
Following its competition with the GIA, the GSPC under Abdelmalek Droukdel’s leadership formally pledged fealty to al-Qaeda in September 2006, changing its name to al-Qaeda in the Islamic Maghreb in January 2007. Indeed, Droukdel’s oath of fealty was the culmination of growing ties between the GSPC and al-Qaeda’s transnational network, particularly al-Zarqawi and al-Qaeda in Iraq.138 Pledging fealty may have offered the GSPC potential sources of financial and logistical support. After all, despite surviving past the GIA’s demise in 2004, the GSPC’s ranks had been diminished in the preceding years due to continued counterterrorism pressure, hundreds of GSPC militants accepting government amnesty programs, and fighters traveling to Iraq.139 Facing such difficulties, Thurston posits that establishing official ties with al-Qaeda may have also constituted “a way to signal to the rank and file not just that the GSPC had international jihadist credibility but also that now there was no turning back.”140 In addition to expressing a desire to attack both “near” and “far” enemies (e.g., France and the United States),141 the newly named al-Qaeda in the Islamic Maghreb launched its first suicide bombing attacks — a tactic that other Algerian armed groups had avoided even during the darkest years of Algeria’s “black decade” of civil war — in April 2007 in Algiers, killing over 30 and injuring more than 220 people.142 The GSPC/al-Qaeda in the Islamic Maghreb’s decision to join al-Qaeda, alter its enemy hierarchies, and adopt an indiscriminate and highly-lethal tactic occurred after, rather than during, the organization’s competition with the GIA. Additionally, by pledging fealty to bin Laden, the GSPC joined al-Qaeda’s ranks prior to the formation of a new al-Qaeda affiliate operating in the Gulf.
Jihadist Competition in Yemen
Al-Qaeda in the Arabian Peninsula (AQAP) formed in January 2009 following the merger of al-Qaeda’s Saudi and Yemeni branches, with al-Qaeda’s Yemeni affiliate absorbing Saudi jihadists who were fleeing from regime security forces.143 Led by Nasir al-Wuhayshi, bin Laden’s former secretary, AQAP quickly developed into one of al-Qaeda’s “most trusted and important affiliates.”144 Shortly after its formation, AQAP expanded its insurgency in the wake of the 2011 Arab uprisings, overtaking territory in various governorates in Yemen’s south. In its attempts to win the “hearts and minds” of local populations,145 AQAP created a front organization — Ansar al-Sharia — to distance itself from any negative connotations associated with the al-Qaeda brand.146 Through Ansar al-Sharia, AQAP delivered long-desired social services and dispute arbitration.147
However, AQAP’s implementation of its interpretation of Islamic law overshadowed its broader outreach campaign. In a similar fashion to the GIA, AQAP attempted to control various forms of social behavior through extreme violence as it controlled territory in 2011 and 2012. For example, the group banned music and dancing, crucified spies, amputated the hands of suspected thieves, and beheaded women who were accused of witchcraft.148 Such measures alienated local civilians and highlighted the negative consequences of such behavior for AQAP. Captured internal al-Qaeda documents highlight that AQAP realized the drawbacks of such violence. Rather than advocating for harsh punishments against violators of Islamic law, al-Wuhayshi noted the need for leniency toward local populations for minor offenses when initially conquering territory:
Generations of the Umma lived without really knowing the tenets of their religion. They have been hard-pressed by the hard toil of making a living. … But Allah has sent you to this Umma to move them toward the right path. You have to be kind to them and make room for compassion and for leniency. …You have to take a gradual approach with them when it comes to their religious practices. You can’t beat people for drinking alcohol when they don’t even know the basics of how to pray. We have to first stop the great sins, and then move gradually to the lesser and lesser ones. When you find someone committing a sin, we have to address the issue by making the right call, and by giving lenient advice first, then by harsh rebuke, and then by force.149
Despite eventually losing control over territory in 2012, AQAP maintained its capacity to conduct large-scale and complex operations throughout Yemen after its failed governing experience. The group used various tactics to target regime personnel, civilians, and rival groups prior to the official al-Qaeda-ISIL split.150
Following this trend, al-Qaeda’s leaders attempted to portray their organization as relatively more moderate in comparison to ISIL.
Outside of Yemen, developments in Iraq and Syria fundamentally altered power structures within the jihadist community over the course of 2013 and 2014. ISIL’s dramatic military triumphs and conquering of mass swaths of territory garnered the admiration of jihadist supporters. Its successes demonstrated, at least to some militants, the utility of its “uncompromising strategy of sectarianism, barbarity, and conquest.”151 Seeking to overtake al-Qaeda’s leadership status among global jihadists, ISIL announced its caliphate in June 2014. In doing so, the newly created Islamic State staked its claim as the sole authority for the global Muslim population while disparaging al-Qaeda chief al-Zawahiri as a “laughingstock,” a sower of dissent among jihadists, and one who had deviated from the rightly guided path of jihad.152 Moreover, ISIL accused al-Qaeda, the once prestigious vanguard of global jihad, of softening its stance at the expense of Islam and adopting a program of action “which believes in pacifism.”153
Indeed, al-Qaeda had attempted to adopt a softer image in the preceding years, pursuing a long-term, gradualist strategy that focused on accounting for local grievances and overlooking doctrinal differences to enhance cooperation with other groups.154 Following this trend, al-Qaeda’s leaders attempted to portray their organization as relatively more moderate in comparison to ISIL. Al-Zawahiri likened ISIL to the Kharijites, an early Islamic sect known for its extremism and excesses in takfir to justify the use of violence against other Muslims.155 Commenting on al-Qaeda’s strategy, Donald Holbrook summarizes: “In presenting the Muslim Brotherhood as being too compromising and weak, and IS [Islamic State] as being too extreme, Zawahiri [charted] a clear ‘middle way’ for al-Qaeda and its understanding of the world, which he claims will ultimately be victorious.”156
Although originating on the battlefields of Syria and Iraq, competition between al-Qaeda and ISIL reached Yemen in late 2014, as the two jihadist powers “engaged in a global confrontation, vying for the hearts and minds of the same constituency.”157 Following the major shock to the power structures of the global jihadist community, AQAP leaders initially adopted a neutral position in the broader power struggle occurring in the bi-polar jihadist landscape between al-Qaeda and ISIL.158 However, AQAP’s position toward ISIL changed in November 2014 after the latter announced a newly created affiliate in Yemen — Islamic State in Yemen (ISY) — and nullification of all other jihadist organizations in the area.
Enter the Islamic State in Yemen
With the Yemeni conflict escalating in 2015, ISY initially expanded through effective recruitment tactics and the defection of AQAP members. For instance, Hadramawt, a province in which AQAP controlled significant territory in 2015 and 2016, became an early recruitment hub for ISY.159 The organization had created multiple “provinces” (wilayat) across various Yemeni provinces, such as al-Bayda, Hadramawt, Lahij, Marib, and Shabwah, by July 2015.160 The establishment of some of these cells occurred in areas in which AQAP became “deeply enmeshed” in the battle against Houthi forces.161 In addition to announcing its presence in different territories, by early 2017 ISY had established two training camps — the Abu Muhammad al-Adnani and Abu Muhammad al-Furqan camps —in al-Bayda in central Yemen,162 an area in which al-Qaeda has historically maintained a presence.163 It is important to note that, despite the influx of foreign fighters to ISIL’s ranks in Syria and Iraq, Elisabeth Kendall notes that “Yemen’s jihad has been largely a local movement,” with the “martyrs” listed in ISY obituaries coming primarily from local areas.164 Facing competition for local support, AQAP needed to effectively handle opposition from a new jihadist rival.
Despite al-Qaeda’s adoption of a “softer image” and AQAP’s recognition of its own mistakes resulting from excess violence, conventional wisdom would expect AQAP to have escalated its violence to more effectively demonstrate its resolve and, consequently, secure the backing of potential supporters during its competition with ISY. Various analysts noted the potential for outbidding among al-Qaeda and ISIL affiliates. For instance, Katherine Zimmerman argued that ISIL’s early successes energized “the entire global jihadist movement, including al Qaeda, to compete with one another in violent conquest and terror.”165 Relatedly, Clint Watts stated that al-Qaeda and ISIL “seemed poised to outpace each other via violence on several continents,” with both organizations’ franchises “aggressively pursuing attacks in an attempt to one up each other.”166 Within Yemen, Jori Breslawski and Brandon Ives contended that, in order to “regain its reign as the leading terrorist organization, Al-Qaeda sought to ‘outbid ISIS on its own field’” and altered its behavior to do so.167 However, rather than seeking to launch more spectacular attacks and lethal attacks, AQAP — in a similar fashion to the GSPC — avoided competing through violent outbidding, seeking to distinguish itself as a more moderate and restrained group than ISY.
Between January 2015 and March 2021, AQAP conducted 921 attacks while ISY conducted 297. However, examining the data on each group’s attacks against Yemeni civilians reveals a sharp difference between the two organizations’ targeting strategies. Though ISY launched only a slightly higher proportion of its total attacks against civilians than AQAP, its attacks were far deadlier. AQAP conducted 94 attacks against civilians, resulting in 136 fatalities during this time frame (1.4 fatalities per attack). In contrast, ISY’s 40 attacks against civilians resulted in 316 fatalities (7.9 fatalities per attack). AQAP’s lower civilian fatality rate was unlikely due to it possessing less capacity than ISY to conduct successful terrorist attacks. AQAP conducted far more total attacks than ISY between 2015 and 2021 and has proven itself capable of successfully launching highly complex armed operations inside and outside of Yemen.168 While possessing the capacity to conduct large-scale attacks, AQAP was much more selective in its use of violence against civilians, avoiding indiscriminate tactics and mass-casualty attacks against civilians.
In contrast to AQAP, ISY’s attacks against civilians consisted of far more indiscriminate bombings, particularly against mosques. As Table 2 demonstrates, ISY committed eight out of the 10 deadliest attacks against civilians recorded by ACLED between January 2015 and March 2021. In March 2015, for example, ISY executed coordinated suicide bombings against mosques linked to Houthi fighters during Friday prayers in Sana’a, killing and injuring nearly 500 civilians. Claiming responsibility for the operation via Twitter, ISY proclaimed, “Let the polytheist Houthis know that the soldiers of the Islamic State will not rest and will not stay still until they extirpate them.”169 Setting an example for future operations, indiscriminate attacks against Yemeni civilians through the use of suicide bombings and remote explosives became a mainstay of ISY’s patterns of violence. The group also demonstrated its propensity for the use of brutality. Shortly after the 2015 Sana’a bombings, a video of ISY’s execution of 14 Yemeni soldiers was released, as the affiliate displayed its adherence to ISIL’s tactic of beheading prisoners. Yet, instead of generating support among the local population by using such violence, the group’s “brutal tactics, including mass killings and mosque bombings … [were] at odds with societal and tribal norms.”170
Table 2: Ten Deadliest Attacks Against Civilians (Yemen)
|ISY||Sanaa-Al Sabeen||73||Suicide Bomber|
|ISY||Sanaa-Al Jiraf||73||Suicide Bomber|
|ISY||Sanaa-Al Jiraf||28||Remote Explosive|
|AQAP||Ahwar||20||Execution of Prisoners|
|ISY||Azzan||14||Execution of Prisoners|
|ISY||Dhi Kalib al Asfal||9||Suicide Bomber|
In addition to differentiating its use of violence, AQAP also rhetorically distanced itself from ISY’s behavior. The theory of outbidding would expect AQAP to claim credit for attacks to bolster perceptions of itself as the most efficient employer of violence in order to win a greater market share of popular support. However, rather than rushing to claim every militant operation, AQAP, in a similar fashion to the GSPC, distanced itself from numerous attacks launched within Yemen. For instance, AQAP denied links to multiple bombings targeting public gatherings (al-Tajammu‘at al-‘Amma).171 This included denying any links to the March 2015 attacks conducted by ISY in Sana’a. At the time, AQAP claimed that it remained committed to al-Zawahiri’s “guidelines” (tawjihat) that advised against targeting “mosques, markets, and public spaces” to preserve Muslim lives.172 In 2016, AQAP also denied having connections to an operation targeting a Catholic-run home for the elderly in Aden that resulted in the deaths of four nuns, stating that this is not the group’s “method of fighting” (tariqatuna fi al-Qital).173
While denying links to disparate acts of violence in Yemen, AQAP also rebuked ISY for its employment of large-scale violence against civilians. For AQAP, such acts were evidence of ISY’s blatant indifference toward civilians’ well-being and its adoption of behavior that had “demolished more than it built and dispersed more than it joined.”174 AQAP argued that extremism was “a scourge” within Islam, a belief system that fundamentally rejected excessive behavior in all matters, including fighting in the path of God.175 AQAP also highlighted ISY’s unrestrained use of takfir to legitimate such violence. According to one ISY defector, ISY leaders accused individuals and entire populations in Yemeni governorates of being apostates (murtaddin) for minor offenses, making them legitimate targets of violence.176
Moreover, AQAP discursively underscored its commitment to being a less extreme group than ISY, helping the al-Qaeda affiliate avoid some of the costs engendered by ISY’s behavior.
Rather than serving Islam, the al-Qaeda affiliate argued that the unreasonably extreme behavior of “al-Baghdadi’s soldiers” served jihadists’ enemies.177 For AQAP, extremism strengthens the interests of Western states, regional “apostate” regimes, and the Houthis, as “the enemy is the greatest beneficiary [al-Mustafid al-Akbar] from the deviation [inhiraf] of Muslims towards extremism.” According to the organization, “innumerable evils and endless calamities” inevitably result from excess behavior. Thus, jihadist organizations engaging in extreme behavior were instrumental in “the destruction of Islam in the name of Islam.”178 Not only had extremist behavior deviated organizations from the rightly guided path of jihad, but the destructive behavior of ISIL and its affiliates had engendered infighting among jihadist groups in important arenas for jihadist activity, such as Afghanistan and Somalia.179
In contrast to the unapologetic actions of ISIL and its affiliates, AQAP publicly apologized for acts of violence that unintentionally killed civilians in attacks targeting the Yemeni military and offered to pay for damages. Moving beyond past transgressions also included rejecting the use of brutal tactics, such as beheadings. When asked about the use of extreme violence that mirrored the behavior of ISIL in December 2014, a senior AQAP leader claimed that “we strongly reject” (nunkiru ‘alayha bi-shidda) scenes of beheadings, claiming that the spread of such acts in the name of religion and jihad is unacceptable.180 Rather than compete with ISY by engaging in such violent tactics or using indiscriminate bombings in public spaces, AQAP unambiguously sought to differentiate itself from ISY and “al-Baghdadi’s Kharijites” by jettisoning past excesses and highlighting the tribulations wrought by extremism.181
The enmity between AQAP and ISY finally came to a head in July 2018, as clashes escalated following ISY’s kidnapping of AQAP members traveling to the frontlines to fight Houthi forces. However, over the course of the prior three and a half years, the organizations had employed disparate forms of violence and bore the consequences of their behavior. While ISY often employed overt brutality and indiscriminate bombings against civilian populations, AQAP showed its relative restraint, strategically using fewer large-scale attacks than its rival. Rather than generating support among the local population, “ISY brutality and indiscriminate attacks alienated Yemenis.”182 Moreover, AQAP discursively underscored its commitment to being a less extreme group than ISY, helping the al-Qaeda affiliate avoid some of the costs engendered by ISY’s behavior. AQAP’s attempts at differentiation were ultimately aided by ISY’s behavior and employment of violence, as “the excessive brutality of ISY gave AQAP the opportunity to look like the ‘acceptable’ face of jihad.”183
There is significant variation in how armed groups compete during conflicts. Understanding armed group competition has become increasingly important, as the number of multiparty civil wars has grown across the world. Violent outbidding is a prominent explanation for how armed groups compete for popular support. This paper outlines an alternative strategy that is available to armed groups facing rivals: strategic restraint. Armed groups may intentionally limit their level of violence during competition to advertise themselves as a moderate alternative and to differentiate themselves from rivals.
We tested the theory of strategic restraint by examining competition between jihadist groups in Algeria and Yemen. According to conventional wisdom, jihadist organizations are “most likely” to employ outbidding. However, rather than escalating their own levels of violence in response to rivals, the GSPC and AQAP juxtaposed their selective violence with the brutality and high casualty rate of the attacks conducted by the GIA and ISY, demonstrating their restraint to local populations. Rhetorically, the GSPC and AQAP advertised their brands by condemning and distancing themselves from a number of highly lethal and spectacular attacks. They also apologized for violent excesses.
This article has several implications for policy. Some scholars and policymakers continue to “lump” jihadist organizations together, masking profound variations among these groups.184 However, our analysis highlights the significant variation that exists in the behavior and discourse of jihadist groups as they compete for public support. Policy directed toward combating these groups will likely be ineffective if these differences are not considered. For instance, state efforts aimed at decapitating group leadership may produce varying results depending on whether an armed group is committed to strategic restraint or violent outbidding. Excessive targeting of a restrained group may simply strengthen more violent competitors. Alternatively, the targeting of more violent groups may facilitate negotiations by strengthening a more restrained organization, or, at a minimum, shift attacks away from civilian targets. Understanding how militant groups strategically compete for popular support through their behavior and discourse thus constitutes an important consideration for debates surrounding leadership targeting.185
Second, our research suggests that how armed groups frame themselves is not cheap talk but instead provides important clues about how they want potential supporters to view them relative to their rivals. Indeed, developing and sharing content online across various platforms remains a prominent way in which jihadist groups reach potential supporters.186 Consequently, counter-messaging, a tool to combat the spread of extremist organizations,187 must account for differences among groups that look similar “on paper.” In addition to accounting for variation in jihadist groups’ discourse, efforts at developing effective counter-messaging may require greater engagement with these groups’ non-English language publications. For instance, a recent study examining articles on ISIL in prominent terrorism journals found that scholarship continues to focus primarily on the organization’s English-language sources, overlooking important Arabic-language materials that ISIL uses to attract support among local populations in the Middle East and North Africa.188
Finally, our research further illustrates the importance of understanding the civilian support base by noting that supporters often possess varying preferences toward violence. If armed groups rely on public support for resources and recruits, then fluctuations in the preferences of potential supporters may significantly impact the strategies of armed groups. Even in the case of a single armed group that has a “monopoly” on violent attacks, that group must consider supporters’ preferences in order to attract maximum backing if it is incapable of ensuring civilian compliance through coercion.
By introducing strategic restraint as an alternative option to violent outbidding, this paper illustrates that the entry of new armed groups into a conflict has important consequences for local populations. If a newly entered group seeks to differentiate itself from pre-existing organizations that rely on brutal tactics, then this newly formed group may adopt lower levels of violence against civilians or even seek negotiations with the state. Further research could explore additional factors, such as ideology, organizational structure, and leadership characteristics, that may lead groups to adopt or maintain a commitment to strategic restraint. Additional scholarship may also identify specific government policies that remove the market share for more violent groups or, at a minimum, shift supporters’ preferences toward favoring less violent groups. As demonstrated in Algeria, there is evidence that governments already update their counterinsurgency strategy to exploit the varied preferences among armed group supporters. Finding ways to measure shifts in the distribution of supporters’ preferences is crucial for identifying opportunities for adjusting counterinsurgency strategies in order to exploit declining support for high levels of violence. Our research suggests that policymakers should identify whether an armed group is pursuing strategic restraint rather than assume any new group will choose to violently outbid existing groups. Attention to such details will continue to be a pressing concern for scholars and policymakers as civil conflicts around the world grow increasingly fragmented.
Mark Berlin is a Ph.D. candidate at the George Washington University. His dissertation research focuses on cooperation between armed groups and variation in their use of extra-lethal violence.
Stephen Rangazas is a Ph.D. candidate at the George Washington University. His research focuses on the effectiveness of state indiscriminate violence during civil wars, primarily using the case study of Vietnam.
Acknowledgments: The authors would like to thank Sultan Alamer, Nick Anderson, Neha Ansari, Leore Ben-Chorin, Nathan Brown, Alex Downes, Dani Gilbert, Charles Glaser, Janet Lewis, Bryce Loidolt, Iris Malone, Rosalie Rubio, Steven Schaaf, participants at the RIP Workshop hosted by ISCS at the George Washington University, and two anonymous reviewers for their excellent and insightful comments.