Spring 2018 catalog cover

The Penn Press list for spring 2018 includes hardcover releases, first-time paperbacks, and ebook editions intended for scholars, students, and serious general readers worldwide. Click here to explore our forthcoming books, grouped by subject area.

Why Terrorist Groups Form International Alliances

Tricia Bacon

May 2018 | 368 pages | Cloth $69.95
Political Science
View main book page

Table of Contents

Chapter 1. A Theory of Alliance Hubs and Alliance Formation
Chapter 2. The Popular Front for the Liberation of Palestine: Pioneering Partnerships
Chapter 3. The Red Army Faction: Pursuing Palestinian Partners
Chapter 4. Al-Qaida Before 9/11: Building Alliances One Dollar at a Time
Chapter 5. Al-Qaida After 9/11: Calling in Debts and Capitalizing on Cachet
Chapter 6. Egyptian Jihadist Groups: Divergent Solutions, Similar Problems


Excerpt [uncorrected, not for citation]


As for the answer to your question, why did we join Al Qaida? We say, why shouldn't we join Al Qaida? God ordered us to be united, to be allied, to cooperate and fight against the idolaters in straight lines. . . . We are a jihadi ancestral community. We rely on legitimacy before anything else as a base of our decisions.
—Abdelmalek Droukdal, leader of al-Qaida of the Islamic Maghreb, formerly the Salafist Group for Preaching and Combat
In 2005, the Algerian Salafist Group for Preaching and Combat (GSPC) was a shadow of its former self. After two amnesties in Algeria and more than a decade of conflict, its ranks had thinned considerably, and its cause of creating a "true" Islamic state in Algeria no longer resonated with the war-weary public. Once an existential threat to the government, the jihadist insurgency was largely a law-and-order problem relegated to the outskirts of the country. Newly radicalized Algerian militants gravitated toward the insurgency against the United States in Iraq rather than join the discredited cause at home. Surrendering members reported that the group's fighters struggled to survive on rations and lived in desolate conditions.

Then, on the five-year anniversary of the 9/11 attacks, al-Qaida anointed the GSPC as its affiliate in the Maghreb. The alliance did not occur overnight. Discussion had been under way for at least a year before the announcement. Al-Qaida had concerns stemming from its acrimonious break with the GSPC's parent group, the Armed Islamic Group. The Armed Islamic Group was reputed to have been infiltrated by the Algerian security services and had alienated even fellow jihadists with its violent excesses. Al-Qaida turned to another ally, its affiliate in Iraq led by Abu Musab al-Zarqawi, which had contacts with the GSPC, for reassurances that the GSPC did not suffer from the same defects. For its part, the GSPC grappled with unity woes about the group's direction and lack of support from the Algerian populace, which culminated in the overthrow of its founding leader. The GSPC's new leader moved to publicly align the group with al-Qaida, signaling his desire for an alliance. Private communications ensued, eventually producing an alliance announcement in late 2006.

In early 2007, the GSPC changed its name to reflect the alliance. The newly minted al-Qaida in the Islamic Maghreb (AQIM) then unleashed renewed violence, shattering the Algerian government's claims that the jihadist insurgency was defeated. AQIM's next attack that March struck a more global target, albeit in the local context, the Russian contractor Stroytransgaz. AQIM also adopted al-Qaida's modus operandi of suicide operations, which broadened its reach and increased the lethality of its attacks. It struck in the heart of Algiers—a locale considered largely secure from the group's bombings and ambushes—not once, but twice. In April, a suicide bomber attacked the headquarters of Algeria's prime minister, killing sixty-seven. Six months later, AQIM struck Algeria's Constitutional Council and the United Nations building with truck bombs, killing over forty people. Simultaneously, the group expanded its lucrative safe haven in the Sahel and attracted some members outside its Algerian base. AQIM had been rejuvenated, and some of the credit went to its alliance with al-Qaida.

The GSPC's transformation reflects how partnering with certain terrorist groups creates opportunities for organizations to improve their resource mobilization and operational capability. Like the destruction produced by AQIM's adoption of suicide operations, terrorist groups with allies, particularly partners that are highly capable, tend to conduct deadlier attacks. As AQIM's renewal suggests, alliances can increase terrorist groups' longevity and resilience. Overall, alliances can improve partnering groups' strength, efficiency, and bargaining leverage.

When alliances occur, they often cluster around select, capable groups like al-Qaida. In essence, hubs operate at the epicenter of alliance networks with numerous satellite groups surrounding them. Consequently, hubs' relationships account for a disproportionate number of alliances. Al-Qaida behaves as an alliance hub, although, as this book will examine, it was by no means the first or last hub. Before al-Qaida, groups like the Popular Front for the Liberation of Palestine, Fatah, and the Red Army Faction also sought to build alliance networks. Moreover, since severing ties with al-Qaida, the so-called Islamic State has emerged as a rival alliance hub. Given the danger posed by hubs' relationships specifically and terrorist alliances more broadly, we need to understand why some groups emerge as hubs and why groups gravitate toward hubs as their partners.

The U.S. government has had little success disrupting or preventing alliance hubs or their networks, even though doing so has been a policy priority for over a decade. As early as 2003, the U.S. counterterrorism strategy asserted that "the interconnected nature of terrorist organizations necessitates that we pursue them across the geographic spectrum to ensure that all linkages between the strong and the weak organizations are broken, leaving each of them isolated, exposed, and vulnerable to defeat." The need to understand alliance hubs shows no sign of dissipating as the Islamic State develops its alliance network and al-Qaida's alliances persevere despite prolonged counterterrorism pressure.

While terrorist organizations garner benefits through alliances, they face substantial obstacles to forging these relationships. As a result, the apex of cooperation—a joint operation involving allied groups—occurred in less than 1 percent of terrorist attacks between 1970 and 2007. This reflects, in part, the hurdles that terrorist groups face when attempting to forge the commitments that alliances require. As the GSPC and al-Qaida's lengthy negotiations suggest, alliances necessitate incurring risks. They involve the possibility of infiltration, betrayal, and additional counterterrorism pressure. They have the potential to alienate followers or cause internal dissension that can lead to splintering. Therefore, terrorist groups must ally carefully and selectively.

Like the GSPC's justification of its relationship with al-Qaida, observers and even terrorists themselves frequently attribute terrorist alliances to shared ideologies and common enemies. Indeed, a strong correlation exists between both shared ideology and alliances and common enemies and alliances. However, the conventional wisdom about ideology and enemies cannot explain why select organizations emerge as desirable alliance partners or how they build their dangerous alliance networks. If enemies and ideology cause alliance hubs, alliances would form where they are often absent. Alone, neither can fully explain the timing of many alliances involving hubs.

Conventional wisdom cannot explain the emergence or appeal of alliance hubs because it omits the influence of organizational considerations. The objectives they seek and the tactics they use can obscure the fact that terrorist groups share core characteristics with all organizations. Fundamentally, like other organizations, terrorist groups seek to survive. More than other organizations, terrorist groups face constant existential danger. Therefore, survival considerations loom large.

Thinly veiled beneath terrorists' declared aims is their belief that they are the ones, perhaps the only ones, who can right the perceived wrongs and precipitate the sought-after change. In other words, terrorist organizations see themselves as indispensable to achieving the change they seek. Consequently, victory depends on organizational survival.

Moreover, not all members join or stay in terrorist organizations because of political or ideological motives. They do not just seek to achieve an organization's declared political agenda; they also value the strong affective ties that come with membership. Consequently, organizational survival becomes an end unto itself, not just a means to accomplishing the aspired-after objectives. The longer a group exists, the more apt it is to substitute organizational perpetuation for its stated goals. This helps explain why groups sometimes persist after their demands are outdated or their motivating grievances are obsolete.

Therefore, rather than shared ideology or common enemies causing hubs to build alliance networks or motivating groups to ally with hubs, I argue that organizational dynamics guide alliance behavior in two main ways. First, hubs become attractive partners because they are highly capable organizations that possess skills, knowledge, and resource mobilization capability that are in demand in the prevailing and anticipated environment. Yet they are able and willing to share these assets in order to address their own organizational needs and ambitions.

Second, for satellite groups, organizational needs precipitate alliance seeking that leads to alliance initiation and formation with hubs. When groups need to acquire new skills, knowledge, or resource mobilization capability, or must adapt existing ones but cannot do so on their own, they seek allies who can help address these needs. Groups tend to experience misalignment with the environment when they are young, recovering from crises, or operating in rapidly changing environments. This organizational weakness and an inability to self-reform prompt groups to seek allies and lead them to hub partners that help realign them with the environment.

However, organizational vulnerability alone is not sufficient for a hub and prospective satellite to form an alliance. When organizational needs arise, groups must be receptive to alliances and to the partners that can assist them. Like all organizations, terrorist groups create cultures, routines, and decision-making processes that enable some behaviors and preclude others, including alliances. Consequently, groups must have organizational processes amenable to alliances in order to consider them as solutions to organizational weakness.

In addition, rather than being the primary cause for hubs to seek allies or why hubs are preferable partners, I argue that shared ideology and common enemies create affinity and guide partner selection, that is, make some partners acceptable. When organizational needs precipitate alliance seeking, groups seek partners that can address those needs and share identity characteristics, particularly ideology or frames that identify the enemies. Hubs fulfill both conditions. They possess identity characteristics, particularly ideology and enemies, which are shared with numerous other groups, making them both acceptable to others and accepting of others. They tend to adhere to the prevalent ideologies of their time and have an expansive view of their enemies. Thus, shared ideologies and common enemies do influence alliances but not in the manner commonly assigned to them.

This introductory chapter explores what we know about terrorist group alliances. It begins by explaining what constitutes an alliance between terrorist groups and how rivalry—or, more specifically, the lack of it—shapes alliance behavior. Next, the chapter examines both the obstacles to alliances and the benefits groups derive from them. It then discusses alliance patterns, namely, the way in which alliances cluster around certain groups, or alliance hubs. Understanding why groups become alliance hubs and their attractiveness as partners are the central inquiries of this book. The chapter concludes with an examination of conventional wisdom and the shortfalls of these explanations in understanding why groups become alliance hubs and why hubs are such desirable partners.

Do Terrorist Groups Ally?

Beyond agreement on its political dimension, which differentiates terrorism from other forms of nonstate violence and crime, terrorism is a notoriously contested concept. By most definitions, terrorism generally consists of three components: political aims and motives; violence or threats of violence against noncombatants or other victims proscribed by laws of war; and intended psychological repercussions beyond the immediate victim or target. As Victor Asal and colleagues argue, terrorism "is in sharp contrast to most other forms of violence, which are largely focused on the immediate target, designed to accomplish a military objective (such as gaining or holding territory), and often at least claim to abide by the codes of conduct eschewing civilian involvement." Terrorist groups—the focus of this book—refer to organizations that use terrorism as "a tactic of violent manipulation." They may engage in insurgency, guerrilla warfare, organized crime, or political activities, but terrorism is a tactic in their arsenal.

A consensus has not yet emerged about what constitutes an alliance. Some have shied away from using the term alliance. Kanisha Bond opted to look at what she termed "inter-violent non-state actor cooperative arrangements," which she defined as a "formal or informal arrangement that has been collectively decided upon by the cooperating parties and governs the management or execution of some level of resource sharing, strategic coordination, and/or tactical collaboration." Ely Karmon chose to use the term coalition rather than alliance because of concerns that alliance assumes too much formality. He defined terrorist coalitions as "ideological, material, and operational cooperation between two or more terrorist organizations directed against a common enemy, which may be a state targeted by one of the member organizations or a rival ideological bloc." Some subsequent work has used elements of Karmon's definition to define alliances. Asal and colleagues drew from Karmon to refer to alliances as "joint or complementary action for the same intermediate purpose. This action can constitute activity at the rhetorical, material or operational level."

It is important to note that terrorist groups can engage in a range of cooperative arrangements, not all of which reach the threshold of an alliance. Assaf Moghadam differentiates "low-end cooperation" from "high-end cooperation" based on the time horizon of cooperation, the level of interdependence, the type of cooperative activity, and the level of affinity among partners. Definitions of alliances that include dyads that engage in low-end transactional or tactical cooperation—groups that maintain full independence and cooperate with a limited time horizon along with those that involve longer time horizons and greater interdependence—risk conflating different types of relationships and thereby obscuring their causes. Consistent with Moghadam's typology, for the purposes of this book, alliances require cooperation involving mutual expectations of some degree of coordination or consultation in the future. This approach encompasses what Bond refers to as cooperative arrangements and departs from other definitions in several ways. First, it requires cooperation, not just complementary action. Second, it does not specify the type of cooperation necessary, whether ideological, material, operational, or other. Instead, it includes cooperation in whatever form it occurs. Third, it does not specify the target or reason for cooperation. This helps to avoid a tautology in which cooperation against a common enemy may be the cause of the alliance and define what constitutes such a relationship. Last, and perhaps most important, unlike the existing definitions, it includes the requirement that alliances involve both cooperation and expectations of future collaboration or consultation. In so doing, it offers a rigorous standard and thereby avoids overestimating terrorist alliance frequency by equating it with cooperation alone.

Cooperation is a fundamental part of alliances. Cooperation involves adjustment and a conscious effort to work together, not simply shared interests or mutually beneficial but uncoordinated behavior. In contrast, harmony refers to a situation in which a group's policies and actions facilitate the attainment of another's goals without coordination. For terrorist organizations, harmony can be easily mistaken for cooperation or vice versa because cooperation can be covert. Groups sometimes engage in acts of harmony, either inadvertently or intentionally. For example, in 2008, the Pakistani terrorist group Lashkar-e-Tayyiba deployed gunmen to Mumbai, terrorizing and paralyzing the Indian megacity for days. The gunmen targeted a train station, hotels, a restaurant, and a Jewish community center, killing over 160 people. Some posited that this attack, particularly striking hotels and a Jewish target, reflected an alliance between Lashkar and al-Qaida. Indeed, al-Qaida embraced the attack as a "heroic" operation. Although Lashkar struck targets that were consistent with al-Qaida's agenda, it did not adjust its plans to accommodate al-Qaida or collaborate with al-Qaida on the attack. Therefore, this was an instance of harmony as opposed to cooperation and thus did not indicate an alliance between the two groups.

In addition, an alliance requires expectations of future consultation or coordination. This criteria reflects Moghadam's argument that high-end cooperation involves interdependence and shared expectations that the relationship will last for an extended period. Some groups that engage in cooperation do not form alliances because they do not develop expectations about future coordination and consultation. In contrast, allies provide assistance to one another that does not require immediate reciprocation because of mutual expectations that their partner will reciprocate when need arises in the future. Like states, terrorist group partners need not specify the degree of commitment or explicitly agree upon specific conditions under which coordination will occur in order to ally. Alliances shape partners' future expectations, even without specific agreements. Notably, alliances do not require—and rarely involve—control. Like other alliances, terrorist partners rarely relinquish all autonomy, exercise control over one another, consult on all matters, or adhere to all of one another's requests.

As noted, alliances do not encompass all forms of terrorist interactions or organizational linkages. Collaboration occurs between members of terrorist groups who are not acting on behalf of their respective organizations. Groups engage in temporary cooperation without expectations of future consultation. They express rhetorical support without actual cooperation or coordination. In practice, it can be difficult to distinguish these types of interactions from alliances. It is not always clear when and if an individual represents his organization. Likewise, given terrorist groups' covert nature, their expectations of future consultation may not be evident to outsiders. Consequently, harmony and individual acts of cooperation risk being conflated with alliances, and the frequency of alliances overestimated, unless the relationships are carefully examined.

Distinguishing Between Rival and Nonrival Alliances

This book examines a subset of alliances that receive considerable attention from the media and counterterrorism efforts: relationships between terrorist organizations that are not rivals. Well-known examples of rivals include Hamas and Fatah or the various factions within the Afghan mujahidin. Rival groups can and do forge alliances, usually cautiously and temporarily. Notably, the consideration that dominates rivals' alliance decisions vis-à-vis one another—relative position—is largely absent among nonrivals. Groups can use gains acquired today against their rivals tomorrow, either directly in armed conflict or indirectly by siphoning off support. Rivals operate in a zero-sum environment: gains endanger one's rivals. Therefore, relative power concerns drive their alliance calculus. In contrast, resources among nonrivals can have a positive-sum quality. When partnering groups are not rivals, they can collaborate without concerns that their partners' gains today will harm them in the future.

What distinguishes rivals from nonrivals? Brian Phillips identified two types of rivals for terrorist organizations: intrafield rivals and interfield rivals. Intrafield rivals are competitors that seek the same primary political goal, such as a Sunni jihadist revolution in the same state or a state for the same ethnic community. In contrast, interfield rivals support conflicting primary goals, such as a left-wing group and a right-wing group, or groups representing different ethnic communities seeking control of the same territory. In some respects, Phillips's work dovetails with William R. Thompson's work on interstate rivalries, specifically arguing that rivals regard each other as competitors, adversaries, and the source of actual or latent threats that pose some possibility of becoming militarized.

Building on Phillips's concept of intrafield rivals, the first type of rival, competitors, are groups that rely on and seek support from the same sources; in other words, they compete in the same political market. Terrorist groups not only attack their enemies; they also seek to crowd out competitors to maintain or increase their political market share. They function in accordance with "competitive exclusion," which holds that groups compete with other organizations that draw upon the same resources. Competitor rivals treat resources as mutually exclusive because they vie for the same finite recruits, funds, or territory. Competitor rivals also seek to improve their status vis-à-vis one another in intangible realms, such as prestige, credibility, and legitimacy, as they seek the allegiance of a common constituency. As Gordon McCormick explained, they "jockey for media time and the attention of a more or less fixed base of potential constituents." In addition to the potential for direct militarized conflict, a competitive dynamic can lead these groups to increase the pace or scope of their attacks to gain support, known as outbidding. Of note, outbidding reflects terrorist groups' tendency to engage in activities driven by organizational considerations rather than strategic objectives, an idea central to the theory proposed in Chapter 1.

In contrast, groups that are not competitor rivals rely on different primary political markets, though there may be some overlap. They tend to focus on different primary foes, conflicts, territories, or political causes. Most important, they do not depend on the same sources for support, recruits, and resources. Therefore, the assets they acquire do not come at the expense of one another. As allies, they can share resources, even personnel. When allies are not competitors, they can have dual members, whereas competitors will be reluctant to share membership. Without common political markets and relative power concerns, less tangible gains, such as prestige or legitimacy, benefit both partnering groups. Noncompetitor allies can simultaneously be the representative of their respective causes without threatening one another's position; improved stature may even enhance a partner's position. To the extent that noncompetitive allies share a cause, it is an overarching one, which helps to unify them without creating rivalry.

However, it is important to note three caveats and refinements to the above discussion of competition. First, a group's competitor rivals are not fixed. Such rivals are determined by a group's political market, which can change. Groups can—and some do—decide to focus on a different adversary or cause. They may decide to relocate or redefine their constituency. The result is a shift in their political market and, by extension, their competitor rivals. Groups that were not initially competitor rivals can become so, and vice versa.

Second, the number of competitor rivals a group faces can fluctuate. Some organizations do not survive, which eliminates competitor rivals. Conversely, new groups form, which adds new competitor rivals. In particular, terrorist groups have a propensity to splinter, thereby producing such rivals. For example, Harakat ul-Jihad Islami, a Pakistani Deobandi militant group, formed during the anti-Soviet jihad in Afghanistan in the 1980s. A few years after its founding, a faction broke away to become Harakat ul-Mujahidin. The two groups later allied, only to subsequently split again. Then, in 1999, Jaish-e-Mohamed splintered from Harakat ul-Mujahidin. One original group eventually resulted in three competitor rivals. All operated in the same political market, vied for recruits from among the same population, sought funds from the same sources, and jockeyed for the attention of the same constituency. Therefore, they competed with one another, and their alliance behavior reflected their preoccupation with their relative position.

Third, groups operating within the same country are not necessarily competitor rivals because national boundaries do not determine competition. This position departs from much of the outbidding literature, which often treats competition as occurring among groups at the domestic level. Instead, I treat competition as a dynamic that can and does occur at various levels, including at the substate or regional level, and thus is not necessarily defined by national boundaries. For example, states can experience multiple conflicts that give rise to multiple terrorist groups that do not share political markets. In addition, states can also serve as safe havens for foreign terrorist groups in places where indigenous groups also operate; these groups may not share political markets, though they reside in the same state.

Pakistan illustrates these dynamics. It faces an internal jihadist insurrection primarily in the Pashtun areas of the country and a separatist conflict in Balochistan. The jihadist groups do not share a political market with the Baloch separatists; therefore, they do not behave as competitors. Numerous foreign jihadist groups find haven in Pakistan. Most of the foreign jihadist groups do not share political markets with local jihadist organizations or the Baloch groups.

Furthermore, groups may compete with organizations that operate in another state. As noted earlier, al-Qaida Core—with its leadership likely based in Pakistan and Afghanistan— competes with the Islamic State, which primarily operates in Syria and Iraq. Therefore, rather than using national boundaries to determine competitors, this book defines competitors and noncompetitors based on whether groups share a political market, irrespective of national boundaries.

However, competitors and noncompetitors are not always neat, mutually exclusive categories because groups can share political markets to varying degrees. Competition operates along a spectrum. On the left side of the spectrum, groups function as competitors: they have substantial overlap in their political markets. On the right side of the spectrum, groups do not behave as competitor rivals, even though there may be some overlap in their political markets. On the ends of the spectrum, the distinction between competitors and noncompetitors is clear. On the far right of the spectrum, that is, in noncompetitive situations, groups have fully distinct political markets. On the far left side of the spectrum, groups have entirely the same political markets.

However, the distinction becomes more complex toward the middle when there is some overlap in political markets, requiring closer examination of the dyad. Until 2014, al-Qaida benefited from being the only Sunni militant group operating in a political market that spanned the Middle East. This regional-level political market limited al-Qaida's competition with fellow Arab Sunni jihadist groups with narrower, national-level political markets. For example, al-Qaida's political market had some overlap with its closest ally, the Egyptian Islamic Jihad, but their interactions fell on the noncompetition side of the spectrum because al-Qaida had a broader political market that included, but did not rely primarily on, Egyptian Islamic Jihad's political market in Egypt. In contrast, the Egyptian Islamic Group and Egyptian Islamic Jihad—both Egyptian Sunni jihadist groups—relied on political markets with substantial overlap. Consequently, the two Egyptian groups competed fiercely with one another but did not behave as competitor rivals toward al-Qaida, as Chapter 6 will discuss.

Sunni jihadist groups with political markets outside of the Middle East experienced even less competition with al-Qaida. For example, al-Qaida's alliance with Jemaah Islamiyah—a jihadist group with a regional-level political market in Southeast Asia—fell further on the right of the spectrum than the Egyptian groups and al-Qaida. Al-Qaida and Jemaah Islamiyah had regional-level political markets in the Middle East and Southeast Asia, respectively, as will be discussed in Chapter 5, and thus fell on the far right of the spectrum.

In 2014, a full competitor rival to al-Qaida emerged. Following the Islamic State's renunciation of al-Qaida and al-Qaida's reciprocal disavowal of it, the Islamic State redefined its political market similarly and broadly enough that it shifted to the far left side of the competition spectrum vis-à-vis al-Qaida. In addition to competing over resources, recruits, and allies, the two groups jockey to lead the Sunni jihadist movement. In the event that al-Qaida and the Islamic State decide to ally again, competitor rival considerations, such as relative position, will guide their calculus.

This book focuses on dyads on the right side of the spectrum. These dyads may have some overlap in their political markets, but they do not rely on the same ones. In so doing, it factors in when competition is relevant to why noncompetitors ally; however, the focus is on alliance seeking, alliance initiation, and alliance formation among nonrivals.

Returning to the criteria for rivals, the second type of rival are adversary rivals, similar to Phillips's interfield rivals. Unlike competitor rivals, terrorist groups that behave as adversary rivals do not share a political market. However, they are also attentive to relative power considerations because an increase in an adversary rival's power poses a potential threat. Adversary rivals typically represent opposing sides of a conflict. This may be a left-wing group's posture toward a right-wing group. Or one organization may operate as a vigilante or an unofficial "paramilitary force" that protects the status quo, while the other seeks to change the status quo. For example, Protestant Loyalists groups in Northern Ireland, like the Ulster Defense Association, were the enemies of the Catholic Republican groups, such as the Irish Republican Army. The two represented different sides in the conflict and behaved as adversary rivals.

Excluding both types of rivals matter because rivals ally selectively, carefully, and often briefly. As Thompson points out, "Dealing with one's rivals entails juggling very real conflicts of interest within a charged context especially prone to various decision-making pathologies (in-group solidarity, out-group hostility, mistrust, misperception, and self-fulfilling prophecies)." Fotini Christia argued that within the context of multiple-party civil wars, which produces abundant rivalries, alliances fluctuate based on groups' assessments of their relative position vis-à-vis one another. The same dynamics appear to function between rival terrorist groups. The rival alliance terrain shifts in response to changes in relative power, such as defections, betrayals, fracturing, detentions, and deaths. Rival alliances thus tend to be tactical and temporary, governed by a desire to acquire the maximum resources as part of the smallest-winning coalition, as long as the risk of exploitation is manageable.

However, notions of relative power and the smallest-winning coalitions do not apply to nonrival conditions. Nonrivals cannot and need not weigh their relative position nor would they create a minimum-winning coalition. Therefore, different considerations govern alliances between nonrivals, a fact that strongly suggests the need for separate theorizing. With two important exceptions, much existing work on terrorist alliances includes both types of relationships or at least does not make an explicit distinction. First, Ely Karmon theorizes about coalitions involving transnational terrorist groups. Second, Michael Horowitz and Philip Potter exclude groups with a common organizational lineage from their analysis. The proposed formulation captures these distinctions but extends them to exclude alliances among all rivals. Limiting the population to transnational groups is insufficient because, as discussed, some transnational terrorist groups behave as rivals. In addition, while groups with shared organizational roots usually compete, rivalry is not limited to them. Therefore, a need exists to further disaggregate rival and nonrival terrorist alliances and then to theorize and test the mechanisms that operate to cause alliances among different types of dyads. Consequently, this book exclusively examines nonrival alliances.

The Challenges to Terrorist Alliances

The attention terrorist alliances receive can obscure how poorly positioned terrorist groups are to form cooperative relationships. Alliances are difficult, dangerous, and a potential source of internal dissension. Terrorist groups must overcome many of the hurdles that states, firms, and their nonviolent counterparts face, but they need to be even more cautious given the abundance of threats to their survival. Simultaneously, they have limited information or access to the strategies that licit entities use to mitigate the risk. Consequently, terrorist groups experience numerous obstacles to alliances.

First, because most terrorist organizations are insular, illicit, and clandestine, they lack the requisite transparency to assure prospective partners that they intend to honor promises and future obligations. Many terrorist groups have short life spans, which means few have the requisite shadow of the future, meaning they expect to interact repeatedly over time, to make cooperation worth the risk. Moreover, terrorist groups struggle to establish reputations as trustworthy partners. Meanwhile, adversaries seek to exploit these fears to undermine alliances and weaken partnering organizations.

Second, and relatedly, terrorist groups often lack sufficient information about one another. In addition to their inherent lack of transparency, terrorist groups can and do misrepresent themselves, hide information, and downplay their weaknesses. Because terrorist groups are secretive about attributes, like their size, strength, finances, and capabilities, prospective partners lack complete, credible information about one another.

Third, terrorist groups do not possess qualities conducive to making credible commitments; moreover, they lack access to solutions to mitigate this hurdle, namely, institutions, mechanisms to punish noncompliance, and third-party enforcement. As it is problematic to enforce accountability on such actors, groups have incentives to use cooperation opportunistically to improve their own security and then defect without reciprocating. Unlike states, terrorist groups cannot overcome commitment hurdles by creating institutions to bind themselves to agreements. Even other nonstate actors, such as transnational nongovernmental organizations, increasingly benefit from institutional forums for coalition building. Nor can terrorist groups enter into contracts enforceable by an outside institution like a firm can. Absent such mechanisms, enforcement between violent nonstate actors is particularly difficult, dangerous, and costly. With few enforcement mechanisms available, fears of free-riding and cheating hinder terrorists' efforts to ally.

Theoretically, the potential exists for a third party, perhaps a state, to enforce cooperation. However, an enforcer must be both able and willing to police partnering groups: a tall order. This requires expending resources to monitor behavior, creating mechanisms to punish noncompliance, and managing reprisals. Presumably, an enforcer would only be willing to undertake this role if the alliance dovetailed closely with its preferences or if it could manipulate the alliance to suit its interests. Furthermore, states sometimes overestimate their ability to manage terrorist clients only to find their expectations exceed their ability. States can help bring groups together, like Sudan under the National Islamic Front or Afghanistan under the Taliban, which will be discussed in Chapter 4. States can also encourage alliances, as Iran did with Hizballah and Hamas. However, these roles differ from acting as an enforcer.

Fourth, allying with another terrorist organization can provoke new enemies and generate additional counterterrorism pressure. Alliances risk inciting groups' adversaries and rivals alike. Like states, one important potential cost of entering into an alliance is that a group may find itself involved in a conflict in which it did not intend or wish to participate. Consequently, terrorist groups may increase the pressure they experience by forming an alliance. For example, al-Qaida opted initially to withhold announcing its alliance with the Somali terrorist group, al-Shabaab, in part for this reason. Usama bin Laden wrote to al-Shabaab, arguing that "if the matter becomes declared and out in the open, it would have the enemies escalate their anger and mobilize against you; this is what happened to the brothers in Iraq or Algeria."

Fifth, organizations risk alienating supporters with their ally choices. Groups' real or perceived constituents range from the international community to local ethnic or religious populations. The potential for constituent disapproval can constrain groups from allying. Or an organization may distance itself from an ally if an ally engages in activities condemned by its supporters. Ethnonationalist groups are arguably most sensitive to this pitfall as they may have clear constituencies, reject the terrorist label, and seek international acceptance. For example, Fatah publicly distanced itself from the Red Brigades in the wake of the international outcry against the Red Brigades' murder of former Italian prime minister Aldo Moro in 1978. At the time, Fatah sought international recognition and could not afford to be associated with the act. Even al-Qaida demonstrated sensitivity to this constraint. In 2005, then al-Qaida deputy Ayman al-Zawahiri wrote to the late al-Qaida in Iraq (AQI) leader, Abu Musab al-Zarqawi, counseling him to modify AQI's behavior, citing the Muslim world's rejection of AQI's beheadings and attacks on Shia religious sites. Al-Qaida feared that AQI's actions were damaging al-Qaida's reputation among Muslims.

Sixth, initiating alliances involves incurring risks for the partnering groups, especially across distances. Because prospective allies may not be proximately located or readily accessible, establishing an alliance involves allocating time, resources, and personnel. Groups may need to deploy personnel or engage in long-distance communications, which are vulnerable to interception and manipulation by adversaries. For this reason, bin Laden expressed concerns about using the Internet for communications and recommended using couriers to ensure secrecy. However, his lieutenants worried that doing so would further inhibit al-Qaida's ability to consult with its affiliates. To cooperate, groups must incur risks, while a failure to communicate increases the likelihood of misunderstandings that can derail alliance formation.

Seventh, alliances risk damaging the internal cohesion of the partnering organizations. Terrorist groups emphasize their exclusivity to help secure their survival, but this hinders outside relationships. Their insularity ensures loyalty and cohesion, but it also discourages connections outside the group. Yet alliances require broadening the circle of identification, involve a loss of autonomy, and have the potential to produce dependence on others. In addition, alliances can cause groups to engage in activities they otherwise would not engage in, which can stoke dissent. Terrorist organizations may not function as unitary actors, particularly given the tradeoffs groups face between operational security versus financial efficiency and operational security versus tactical control. Therefore, alliances risk exacerbating tensions and provoking internal dissension, even causing splintering. In addition to the damage splintering does to the original organization, offshoots become rivals.

Eighth, alliances involve dealing with the egos and idiosyncrasies of covert and suspicious individuals. In particular, terrorist leaders can be charismatic figures with an inflated sense of purpose and ambition. Cooperation and compromise between such personalities may be problematic, to say the least. Yet, as will be discussed in Chapter 1, leaders play a pivotal role in alliance formation. As a result, personality clashes can stymie alliances.

Ninth, cultural and political differences may hinder alliances. Nonrival groups in particular often hail from different cultural and political backgrounds, which increases the possibility of misunderstandings and disagreements. Cultural clashes contributed to tensions between the Red Army Faction and Fatah when the West Germans insisted on mixed-gender sleeping accommodations and engaged in sunbathing behavior that offended their conservative Arab hosts, as will be discussed in Chapter 3.

Last, and of the greatest consequence to terrorist organizations, alliances increase the possibility of betrayal. As clandestine organizations operating at a material disadvantage to their adversaries, terrorist groups must be vigilant to prevent enemy infiltration. Alliances increase the number of actors with knowledge of a group's activities, thereby increasing the possibility of security breaches. Allies are a potential source of deliberate or inadvertent security lapses. Groups struggle to verify that their partners have not been compromised and are adequately security conscious; therefore, alliances involve incurring the serious danger of infiltration.

Therefore, terrorist groups are cautious about alliances. They must overcome numerous obstacles and be willing to incur risks. Nevertheless, a small number of groups build alliance networks and emerge as hubs within the terrorist alliance landscape with devastating consequences. Far from a natural outcome, this presents a puzzle. This book seeks to illuminate why alliance hubs opt to build alliance networks, how they do so, and what motivates groups to ally with them.

What Can Groups Gain from Alliances?

What could groups acquire through an alliance that would warrant undertaking such risks? Indeed, terrorist groups can derive an array of benefits from alliances. The advantages range from material or reputational gains, to knowledge acquisition, to increased political and organizational skill. Through alliances, groups can maximize their efficiency by leveraging one another's specializations and comparative advantages. Under some conditions, they may even aggregate their capabilities.

An ally may improve its partner's ability to acquire various kinds of assets. This includes materiel, such as weapons or equipment. For instance, the Red Army Faction reached out to Fatah in 1970 in order to access firearms not readily available in West Germany, which led to their alliance attempt. Under some circumstances, alliances offer improved credibility, legitimacy, and prestige, which can translate into assets like recruits or funds. Alliances can generate propaganda, which may also lead to more recruits and funds. Overall, alliances can help groups improve their resource mobilization capability.

Groups can sometimes project themselves as part of a broader movement or another cause through alliances, a particularly valuable asset when the resonance of their cause wanes. By incorporating an ally's cause, groups can improve their framing resonance. Al-Qaida offered its allies a way to shift or expand their frames from revolutionary jihad focused on overthrowing the near enemy, that is, national governments, to global jihad against the United States, which improved some groups' framing resonance after their originating cause lost its appeal. For example, the GSPC acquired this benefit when it allied with al-Qaida to become AQIM. Though few recruits were attracted to the cause in Algeria, more sought to join an organization associated with al-Qaida's cause after 9/11.

Alliances offer avenues for groups to acquire new capability, expand existing capacity, or conduct ongoing activities more efficiently. Groups can provide one another with knowledge, train one another in new skills, or teach one another improved techniques. Through alliances, groups can share best practices and lessons learned, thereby improving and accelerating one another's knowledge, skill, and materiel acquisition processes. An organization can learn more effective methods and avoid mistakes made by their partner. An organization's research and development can benefit its partners as well. For example, al-Qaida advised the AQIM on how to avoid aerial surveillance based on its experience in Afghanistan and Pakistan. Or allies may recommend shortcuts that produce the same outcome using less manpower or equipment.

While alliances may precipitate additional counterterrorism pressure, conversely, alliances can help groups withstand or circumvent counterterrorism pressure. In particular, alliances help groups survive in hostile environments, when operating in states with greater capacity or in places with autocratic regimes. Furthermore, alliances can diffuse counterterrorism pressure among more groups, thereby reducing the amount that individual groups experience. For this reason, al-Qaida encouraged its allies to attack in the West, noting in its internal communications that "it is better for us that someone will share this responsibility with us and also to disperse and scatter the enemies efforts, instead of concentrating only on Al-Qaida." Groups that enjoy sanctuary may offer partners a place to operate beyond their adversaries' reach. In addition, allies may provide facilitation or logistical assistance that makes their activities less detectable.

Alliances can improve partnering groups' positions vis-à-vis their competitors and their bargaining position. An ally may help a group to compete more effectively with its rivals. Few groups are immune to concerns about competition because of terrorist groups' propensity to splinter. However, an ally can help a group to differentiate itself from rivals or attract more support and resources. Following the Islamic State's announcement that it had formed a caliphate, a number of the initial pledges of allegiance came from small groups seeking to improve their relative position in their political market by aligning with a partner with prestige in jihadi circles. In addition, signals of outside support can enhance groups' bargaining strength.

Alliances offer a number of benefits, and, under some conditions, they are sufficient to warrant the risks. For some groups, the assets acquired through an alliance can be the difference between survival and extinction or between victory and defeat. Given groups' propensity to ally with select, highly capable organizations, it appears that certain groups—alliance hubs—are better positioned than most to bestow these benefits.

The Groups at the Center of Alliance Networks

Given the risks and potential gains involved, terrorist groups select their partners carefully. Following the internationalization of terrorism in 1968, cooperation between terrorist actors flourished to previously unseen levels. At the time, a controversial theory gained traction, positing that this increased collaboration was state driven, a conspiracy hatched by the Soviet Union and the Communist bloc to coordinate terrorist groups' efforts against the West. Instead, Kent Layne Oots found that a small number of terrorist groups acted in a manner akin to "major powers" by providing aid and resources to others, primarily those of similar ideological orientation. Similarly, in his examination of al-Qaida, Jason Burke likened its alliance behavior within the Sunni Islamic militant milieu to a great power's alliance posture in the international system. More recently, Horowitz and Potter similarly found that "selective cooperation results in a core/periphery structure in the network of terrorist alliances." Asal and colleagues also found that "terrorist organizations tend to prefer network structures that are organized into cliques or subgroups." Clearly groups at the center of these clusters, termed alliance hubs, are desirable partners. Not surprisingly, they also tend to be highly capable organizations.

An alliance hub is a group that operates at the center of an alliance network. Alliance hubs function as focal points around which groups orbit and to which they are drawn. Like states, terrorist organizations seeking alliances with the most advantageous partners can lead to a system in which a few groups become highly connected. This clustering around certain entities resembles what network scientists call a star: one entity at the center with ties with other entities radiating from it. Hubs demonstrate an exceptional propensity to ally with other organizations. They possess an ability to more readily overcome the obstacles to cooperation and forge alliances, thereby amassing and distributing the corresponding benefits.

A growing body of research indicates that hubs' alliances have dire consequences. In general, research on social movements indicates that organizations at the center of alliance networks garner the most benefit in terms of knowledge diffusion and exposure to novel tactics; they are also the most active exporter of tactics. Alliance hubs build terrorist coalitions that influence and shape broader terrorist movements. They have agenda-setting power and influence the behavior and objectives of allies. More specifically, Horowitz and Potter found that terrorist groups closely allied with hubs are more likely to gain capability that leads to greater lethality. They also found that connections to hubs improve groups' capability more than alliances with less-connected organizations. In addition, Phillips concluded that multiple alliances improve groups' longevity; therefore, hubs become more resilient through their alliances. Consequently, understanding why groups become alliance hubs and why they attract partners will lead to more effective policy interventions.

The existence of alliance hubs poses a puzzle in light of the obstacles. And their alliance success is anomalous. Hubs' exceptionality—and impact—makes them both important and theoretically difficult to explain. This leads to the central questions of this book: Why do terrorist group alliances cluster around hubs? What makes hubs desirable alliance partners? Why do hubs seek to build such coalitions?

The Shortfalls of Conventional Wisdom in Explaining Alliance Hubs

The factors that motivate groups to become alliance hubs and other groups to cluster around them remain simultaneously undertheorized and overgeneralized. The prevailing wisdom on terrorist alliances assumes that common enemies or shared ideologies motivate alliances, while remaining silent about the many instances when these commonalities exist, but do not produce alliances with a hub. Moreover, the causal mechanisms by which ideology or enemies produce an alliance with a hub are underspecified. What constitutes shared ideologies or enemies is sometimes applied in an elastic way, fitted to explain the presence or absence of a partnership on a case-by-case basis with little consistency or predictive accuracy. It remains unclear how common enemies should be weighed or what level of ideological compatibility causes an alliance.

The shortfalls of conventional wisdom become more acute when examining the alliance patterns that produce hubs. If ideological solidarity causes alliances with hubs, why do alliances cluster in a spoke-and-hub arrangement? Conventional wisdom does not provide insight into why the spokes do not ally equally with one another, as groups in the cluster would seemingly share an ideology as well. Similarly, common enemies do not fully explain this alliance pattern because spokes should also oppose the same adversaries and thus should equally ally with one another as well.

Is the Enemy of My Enemy My Friend?

Governments, terrorist groups, experts, and even casual observers often invoke the adage that terrorist organizations ally to fight common enemies. Indeed, groups with a shared target do have a higher probability of working together. This idea is used to explain "strange bedfellows," groups that have ideological differences but cooperate against shared enemies. For example, in the indictment of the 1998 bombings of the U.S. embassies in Kenya and Tanzania, the U.S. government charged that "bin Laden . . . and other ranking members stated privately that al-Qaida should put aside its differences with the Shiite Muslim terrorist organizations, including the Government of Iran and its affiliated terrorist group, Hizballah, to cooperate against the perceived common enemy, the United States and its allies." Similarly, commenting on the prospects for a relationship between al-Qaida and Hizballah, counterterrorism expert Paul Pillar remarked that given the two groups' common enemies, he "would expect, even with all the distrust and rivalry, they would see ways in which they could cooperate."

Terrorist groups themselves, including hubs and their partners, publicly invoke common enemies to explain their alliance decisions. Even underlying bin Laden's numerous exhortations to fellow Sunni groups to ally with al-Qaida was the argument that unity was necessary, given the shared threat they faced from the "Crusader-Zionist" alliance. Similarly, in its declaration announcing the alliance with al-Qaida, the GSPC invoked this rationale.

The destruction of war, the difficulty of the present situation, and the unified coalition of our enemies against us make it necessary for us to confront this coalition with our own coalition, their alliance with our alliance, face their unified forces against our unified forces. . . . The United States of America will only be defeated by a United States of Islam. . . . Our brothers in Al-Qaida organization under the lead of Mujahid Shaykh Usama Bin Laden—may Allah protect him—are the best ones in this era to unify the scattered Muslims against their enemies.
Or Brothers-in-Arms?

Ideological solidarity is another oft-cited explanation, frequently offered after an alliance occurs between groups of the same ideological disposition. Indeed, alliances tend to involve groups with ideological commonalities. Though rarely explicitly explained, the idea behind ideological solidarity seems to be that terrorist groups with common ideological platforms are attracted to one another owing to their similarities.

Terrorist groups often explain their alliance decisions as necessary acts of ideological solidarity. When asked why the GSPC decided to ally with al-Qaida, the GSPC's leader invoked ideological solidarity as well: "Joining was a legitimate necessity by the book of our God and the sunnah of our prophet, peace and blessing be upon him. . . . Many analysts and observers are mistaken when they think that our joining was a result of secular accounts and self-interests. We are a jihadi ancestral community. We rely on legitimacy (from religion) before anything else as a base of our decisions."

The Deficiencies of Conventional Wisdom

However, the idea that common enemies and ideological solidarity explain the phenomenon under examination—alliance hubs and the nonrival groups that ally with them—faces a number of shortcomings. First, while hubs have multiple alliances, both predict that alliances should form with hubs where, in many cases, none exists. Therefore, it is unclear when, and even if, shared ideology or common enemies are sufficient to motivate an alliance with a hub. Taken alone, shared ideology and common enemies are indeterminate; groups with comparable levels of shared ideologies or enemies have different alliance outcomes vis-à-vis hubs. They struggle to explain why some alliance efforts with hubs fail while others succeed.

Second, these explanations are often invoked post hoc. This opens the possibility that dyads involving alliance hubs may have more common enemies—and experience the accompanying threat—or appear to have greater ideological convergence as a result of an alliance rather than as the cause of it.

Third, enemies and ideology do not often shift significantly during most groups' life span, and so they cannot explain the timing of an alliance with a hub. For example, al-Qaida and Hizballah cooperated temporarily in the 1990s, though they continued to oppose common enemies after their cooperation subsided. Likewise, the GSPC and al-Qaida shared an ideology and enemies for years before "God ordered them to unite."

Fourth, ideology and enemies are sufficiently malleable and underspecified—that is, groups have numerous ideological tenets and enemies—that terrorist organizations can easily cherry-pick and claim that they allied because of ideological tenet X or enemy Z when neither X nor Z actually drove their alliance decisions.

Fifth, many terrorist groups have multiple enemies, even if most primarily focus on one adversary. Yet they only ally with a subset of the groups that share these enemies. Al-Qaida opposes the West, particularly the United States and Western European countries, as well as so-called apostate regimes. The Islamic State opposes virtually everyone who is not part of the group, even other Sunni Muslims. In other words, they share enemies with many groups operating but have only allied with a subset of them.

Despite the commonsense appeal and parsimony of conventional wisdom to explain nonrival alliances involving hubs, these shortfalls suggest that at a minimum there is an omitted causal mechanism operating. Furthermore, conventional wisdom may mis-specify how shared ideologies or common enemies operate in the alliance process. As this book seeks to demonstrate, alone or in tandem neither can fully explain alliance hubs or their partnerships. Instead, I will argue in Chapter 1 that enemies and ideology do play an important role in alliances with hubs by influencing partner selection and bolstering affinity. In particular, alliance hubs' expansive ideologies and enemies make them acceptable partners to numerous organizations. Rather than reflecting their primary motives, groups may invoke shared ideologies and common enemies to mask the organizational weakness that drives their alliances.


Alliance hubs and their partnerships will continue to heighten the threat that terrorist groups pose. In order to more effectively divide terrorist organizations and prevent future alliances, counterterrorism officials must better understand what makes alliance hubs such compelling alliance partners and how they are able to build alliance networks. While al-Qaida has often been viewed as a unique case, other terrorist organizations have forged and benefited from extensive alliance networks. As Chapter 2 will examine, militant Palestinian organizations, especially the Popular Front for the Liberation of Palestine and Fatah, cultivated extensive ties with other terrorist organizations in the late 1960s and 1970s, serving as alliance-hub pioneers. In the 1980s, the Red Army Faction endeavored to build a leftist Euro-terrorist alliance coalition. More recently, the Islamic State has emerged as a rival alliance hub to al-Qaida. Therefore, conditions and motives exist that span time, geography, ideology, and international conditions to produce constellations of alliances revolving around hubs. Meanwhile, existing explanations fail to convincingly explain or predict this aspect of terrorist alliance behavior. Instead, we must look more closely at the organizations themselves.

The goal of this book is to further our understanding of several interrelated questions about this process: (1) What causes groups to seek an alliance despite the risks? (2) Under what conditions do groups select hubs as partners or agree to ally with hubs? (3) When does alliance initiation lead to alliance formation? (4) Why do alliance hubs opt to build alliance networks?

Chapter 1 begins that endeavor by offering a theoretical framework that blends organizational needs, identity affinity, and trust. First, organizational weakness prompts alliance seeking or receptivity to hubs. Hubs become desirable partners because they are positioned and willing to address others' organizational deficiencies. Organizational needs interact with identity affinity, including ideology and common frames about the enemy, to guide partner preferences and create a sense of affinity. Hubs possess identity features acceptable to numerous organizations and accept groups with shared identity features. Last, partners must build trust, if it does not already exist. Hubs develop the ability to forge trust, particularly through brokers who are adept at building interorganizational trust. Using a wide variety of evidence and drawing from multiple disciplines, Chapter 1 lays out the conditions that lead to alliance formation with a hub:

  1. the emergence of acute and/or growing shortfalls in a group's knowledge, skills, or resource mobilization capability as a result of organizational youth, organizational crises, or unanticipated changes in the environment;

  2. a recognition within the group, especially by leaders, of these deficits, an inability to address them through self-reform, and internal dynamics that are receptive to alliances;

  3. the identification of a hub partner that can help address the organizational deficits and that shares identity characteristics, particularly ideology and/or enemy narratives; and

  4. a willingness to build trust based on shared identity characteristics, previous interactions, personal friendships, and/or reputation.

Chapter 1 concludes with a discussion of the book's methodology, including case selection, and lays out the plan of the remainder of the book.