Abstract
Militias are an empirical phenomenon that has been overlooked by current research on civil war. Yet, it is a phenomenon that is crucial for understanding political violence, civil war, post-conflict politics, and authoritarianism. Militias or paramilitaries are armed groups that operate alongside regular security forces or work independently of the state to shield the local population from insurgents. We review existing uses of the term, explore the range of empirical manifestations of militias, and highlight recent findings, including those supplied by the articles in this special issue. We focus on areas where the recognition of the importance of militias challenges and complements current theories of civil war. We conclude by introducing a research agenda advocating the integrated study of militias and rebel groups.
Keywords
Most studies tend to approach civil wars as bilateral conflicts, that is, the state on one side versus the rebels on the opposite side. While recent research has challenged this view by pointing to extensive fragmentation of the rebel side (Bakke, Cunningham, and Seymour 2012), researchers have tended to disregard the proliferation of armed actors that emerge during an armed conflict to fight on behalf of the state and/or against the rebels. Such “paramilitaries,” “civil militias,” or “civil defense forces” are armed groups that operate alongside state security forces or independently of the state, aiming to shield local populations from rebel demands or depredations and seeking to acquire its loyalty or collaboration. Examples include the paramilitaries in Colombia, the hunter militias and civil defense forces (Kamajors) in Sierra Leone, and the comités de autodefensa in Peru. More recent examples featured prominently in international headlines are the Shabbiha militias supporting the Assad regime in their fight against the insurgents in Syria, the Civilian Joint Task Force—anti–Boko Haram militias—in Nigeria, and the units organized to combat pro–Russian separatists in Ukraine. Whatever their name and origin, these forces have a multilayered and significant impact on civil war dynamics. Their emergence affects the process of armed group recruitment, the levels and patterns of violence against civilians, the dynamics and duration of the conflict, and the modalities of its termination; they can also shape the conflict’s outcome and the prospects for long-term post-conflict stability.
Militias are common in civil wars. Carey, Mitchell, and Lowe (2013) find that pro-government militias, which only constitute a subset of this type of armed group, were present in 81 percent of all country-years affected by civil war between 1981 and 2007. Although militias emerge in most conflicts, rarely are they part of comparative studies of civil wars, be they empirical or theoretical (but see exceptions, Humphreys and Weinstein 2008; Gutiérrez Sanín 2008; Arjona and Kalyvas 2012). Recent studies have underlined the fragmented nature of rebel groups in civil wars (Pearlman 2009; Pearlman and Cunningham 2012; Cunningham, Bakke, and Seymour 2012; Staniland 2012; Kenny 2010; Christia 2012) and have called for a revision of the usual dichotomy between state actors and insurgents (Cunningham, Gleditsch, and Salehyan 2009). Yet they remain agnostic about similar fragmentation processes on the side of the state. Indeed, the state is still widely perceived as a unitary actor (Carey, Mitchell, and Lowe 2013, 250; Mazzei 2009, 6).
The articles assembled here aim to advance our understanding of the formation, organization, and behavior of militias in civil wars. Unlike studies that label all types of nonstate armed groups “militias,” including insurgent-affiliated armed groups, we focus exclusively on actors fighting against insurgents. 1 Unlike other definitions (Carey, Mitchell, and Lowe 2013), we do not require a recognizable link to the state to define an armed group as a militia. Militias are often controlled or co-opted by government representatives, but they can shift their loyalties and may pursue agendas that are at odds with the interests of the state. The critical issue, we argue, is their anti-rebel dimension. 2
We begin by exploring the range of empirical manifestations assumed by militias, highlighting recent findings, including those by articles in this special issue. We focus on dimensions where the recognition of the presence and role of militias challenges and complements current theories of civil war, such as armed group formation and recruitment, the transformation of wartime social networks and institutions, and the dynamics of violence. We also discuss how the study of militias contributes to broader questions, such as the impact of militia activity on the state’s monopoly on the use of force, the trajectory of state and nation building, and the quality of new democracies. Explaining the relationship between states and militias and the interaction between rebel groups and militias are promising avenues for future research. Finally, we argue that the contributions of militias to a variety of conflict processes have been systematically overlooked in previous research and that the integration of these actors into general theories of armed conflict can improve our understanding of the duration and termination of civil war as well as their internal dynamics.
Occurrence and Variation
Militias operate in both democratic and authoritarian states as well as during times of peace and war (Carey, Mitchell, and Lowe 2013). They have been used extensively as auxiliary troops of regular armies in interstate wars. Carl von Clausewitz was a fervent advocate of the use of popular militias, both in an offensive capacity on foreign soil and defensively in case of foreign invasion. Their two key features were “a very extensive more or less voluntary co-operation of the whole mass of the people” and their auxiliary role to regular forces (Clausewitz 1976, Book VI, Chapter VI, 372). The American Constitution’s Second Amendment connecting “the right of people to bear arms” with the presence of a “well regulated Militia, being necessary to the security of a free State” is perhaps the best known reference to this type of military organization.
Existing research on militias is mostly limited to case studies, while a few studies on armed groups incorporate militias. Comparative studies examining different dimensions of variation are rare and vary with respect to their definition of militias, data sources, and methods of analysis, so that their results are difficult to compare. This is partly due to a lack of systematic data on militias, although recently there have been important advances in that regard. Carey, Mitchell, and Lowe’s (2013) cross-national Pro-Government Militias Database contains information on all militias with a clear link to the state from 1981 to 2007. 3 Furthermore, subnational studies have compiled novel data on militia activity in specific contexts (e.g., Eck 2015, on Burma).
Militias have emerged in all kinds of civil wars, including irregular civil wars in which states seek high-quality local information (Kalyvas 2006), ethnic or separatist insurgencies (Kalyvas 2008b), wars against foreign occupiers (Branch 2007) as well as conventional and symmetric nonconventional civil wars (Kalyvas and Balcells 2010). They vary with regard to size, location of operation, level of professionalism, and tasks; formation, recruitment, and membership; their relation to the population and their use of violence; and their relation to the state.
One of the most obvious distinctions is size (Kalyvas and Arjona 2005). Militias can grow into large armed organizations or remain relatively small. They may operate locally or spread geographically by becoming mobile (Kalyvas and Arjona 2005; Jentzsch 2014; Fumerton 2002). They can evolve into quasi-professional armies with trained soldiers using professional equipment, such as Colombia’s paramilitaries, or remain ad hoc formations of largely untrained villagers using machetes, bows, and arrows, as in the case of the anti–Boko Haram militias in Nigeria. Both types can in fact coexist and partake in a division of labor. In Mozambique, mobile, professionalized forces were more often involved in offensive military tasks, while the activities of local forces typically remained limited to defense activities such as nightly patrols (Jentzsch 2014).
Like the members of rebel groups, militia members may be recruited coercively or join voluntarily (Cohen and Nordås 2015). To recruit members, militias may rely on existing social networks or create new networks for mobilization (Stanton 2015; Forney 2015). They may target for recruitment defectors from rebel groups, or civilians already living under (partial) territorial control of the government (Kalyvas 2008b). The interaction between rebels and militias in various military, political, and social arenas is crucial for understanding the process of militia recruitment.
Finally, the militias’ relation to the population varies both in terms of the level of violence perpetrated against civilians and in the provision of public goods to the community in which they operate. While some militias start out as primary protectors of civilians, they may later engage in predatory behavior. The Civil Defense Forces in Sierra Leone, for example, initially formed to protect the population against attacks by the Revolutionary United Front rebels, were among the main perpetrators of violence against civilians during the country’s civil war (Sierra Leone Truth and Reconciliation Commission 2005). Sometimes, predatory behavior against civilians is limited to those suspected of working with rebels, but militias may also take advantage of their powerful presence in a village and coerce or abuse civilians they are supposed to protect (Forney 2015). States may also use militias as a way to delegate certain types of violence against civilians (Carey, Colaresi, and Mitchell 2015; but see Cohen and Nodås forthcoming). Some militias provide public services in addition to security. One of the least explored aspects of militias is their role in setting-up local forms of governance that depart from existing state structures and can operate with remarkable autonomy (García-Godos 2006; Schubiger 2013a; Schubiger 2013b). Like rebel governance (Arjona 2010; Mampilly 2011), militia governance can vary considerably as a response to group strategies and local conditions.
Militias and Their Relation to the State
Overall, a key distinction between militia organizations hinges on their relation to the state. The incentives of state actors to form and support militias differ from the dynamics that drive community-based militia formation and mobilization (Jentzsch 2014; Schubiger 2013a). This distinction is crucial if we are to understand where, when, and why militias form, mobilize, and fight. Therefore, researchers should be explicit about the assumptions they make about the relationship between local communities and the state. The distinction between state-orchestrated and community-driven forms of militia formation and mobilization, however, does not imply that these patterns are stable over time. Militias that were formed by communities may later receive assistance from state agents to mobilize recruits. Likewise, state-imposed militias may organize themselves and mobilize followers independently of the state at later stages (Jentzsch 2014; Schubiger 2013a).
Recent studies have highlighted the state’s delegation of violence to militias in civil wars, as states have particular incentives to set up or collaborate with militias during armed conflict. First, militias may be a feasible and effective means of counterinsurgency, providing valuable local knowledge and efficient means to collect intelligence (Kalyvas 2006, 107; Kalyvas and Arjona 2005; Branch 2007; Lyall 2010). In particular, extensive resources and a dire need for local collaboration led imperial, colonial, and foreign occupying powers to rely extensively on local militias (Kalyvas 2008b). Second, militias may serve as force multiplier for regular armed forces, especially when they are weak. For example, after military purges, states may make use of militias for counterinsurgency purposes in regions where they face domestic rebellions and instability (Eck 2015). Third, weak democracies and recipients of financial aid from other democracies may rely on militias to avoid being held accountable for the perpetration of violence against civilians, an instance of plausible deniability (Carey, Colaresi, and Mitchell 2015). Fourth, a regime’s ideology may influence the state’s strategies toward militias. While ideologies emphasizing regime centralization, such as Marxism–Leninism, tend to facilitate collaboration with, and strict control of militias, more fragmented regimes tolerate the use of militias by political elites for parochial or private purposes (Staniland 2015). Finally, authoritarian regimes may create militias as a method of “coup-proofing” (De Bruin 2014, 3-4).
A common error is the assumption that militia assistance to the state denotes a relation of full subservience—in other words, the claim that militias are mere “puppets.” However, a state’s strategic collaboration with or tolerance of militias does not mean that it necessarily has complete control over their formation and activities. As discussed previously, militias may also form independently of state policy, so as to provide security and protect property. During civil wars, communities may choose to protect themselves against wartime violence by forming militias, independently of state forces (e.g., Blocq 2014; Schubiger 2013b; Jentzsch 2014). Even in cases in which states impose or co-opt militias and influence their activities, these groups may evolve into forces with their own goals and interests, such as in the case of the paramilitaries in Colombia or the Civil Defense Forces in Sierra Leone (Acemoglu, Robinson, and Santos 2009; Chaves and Robinson 2010).
Implications for Theories on Armed Conflict
Research on militias carries significant implications for theories of armed conflict in general and civil war in particular. Taking militia activity into account can alter our perspective on civil war duration and termination, civil war violence, the transformation of wartime social networks and institutions, and recruitment for armed groups.
Duration and Termination
Militia activity may increase the length of civil wars. Previous research suggests that civil wars with multiple insurgent factions last longer, as veto players (Cunningham 2006), spoilers (Pearlman 2009) as well as information asymmetries and commitment problems among the various factions (Walter 2009) can prevent reaching a settlement. Militia activity during civil wars is a neglected variable in these studies, as militias may severely complicate the negotiation of settlements. Governments will not only have to come to terms with opposing rebel groups during settlements but also meet the expectations of paramilitary groups. At the same time, the likelihood of conflicts ending in decisive government victories may be influenced by the presence of militias. Peic (2014) reports that governments are 53 percent more likely to preserve their authority if the incumbent deploys a civil defense force. An example is the war in Iraq, where the alliance between Sunni militias and the US military helped turn the tide against the insurgents, albeit only for a short time period.
The dynamics of rebel–militia interactions are likely to influence the dynamics and outcome of civil wars. Because militias recruit locally and undercut the rebels at the grassroots, the latter often perceive militias to be their primary enemy. In Nigeria, for example, the activities of the Civilian Joint Task Force have led to reprisals by Boko Haram against a community in which the militia has been active (Human Rights Watch News 2013). Rebels target militias, but this may backfire and cause the opposite result, an explosive growth of militias, which may contribute to the decisive defeat of the rebels (Degregori 1998). Conversely, when militias turn predatory and indiscriminate, they may fan the growth of the rebels (see Branch and Wood 2010, 6-7). A fine-grained understanding of the social and political context aides the study of these interactions, as militias also provide violent means to settle local conflicts (Kalyvas 2006).
Clearly, the consequences of militia activity on processes of conflict duration and termination require more research. A way forward is to disaggregate militias in terms of their actual function and the context in which they operate. One hypothesis here is that the creation of auxiliary militias by stronger states may shorten conflict duration and promote a favorable outcome for the incumbent, whereas the emergence of more autonomous militias in the context of weak states will increase the duration of conflicts and reduce the probability of insurgent defeat. Precisely because certain types of militias (or certain uses of militias) may be causally associated with various outcomes at the micro and macro levels, we need to identify the determinants of variation in the type and use of militias.
Civil War Violence as Triggers of Militia Mobilization
There are two important dimensions of the relationship between civil war violence and militias. First, civil war violence—whether insurgent or incumbent—can contribute to the formation of militias. Second, militia activity may reduce or increase the level of violence or change the types of violence observed.
As outlined previously, militias can be established from above, by state agents or grow from below, out of local initiatives. The latter often occurs as a reaction of civilian communities to insurgent violence, at times conditional on local dynamics and social context. For instance, based on fieldwork in South Sudan, Blocq (2014) argues that local armed self-defense forces emerged as a reaction to indiscriminate insurgent violence only if affected communities interpreted insurgent violence as originating from neighboring tribes, a condition under which armed resistance was perceived both necessary and manageable. This is an example of how studies on militias contribute to the notion that civilians are active participants rather than passive victims of war, as they draw attention to civilian agency in the context of complex local interactions.
Counterintuitively, the formation and mobilization of militias may also be triggered and reinforced through exposure to indiscriminate state violence. Lyall (2009) argues that civilian counterinsurgent resistance may be one of the key mechanisms through which exposure to shelling attacks by state forces decreased subsequent insurgent violence in Chechnya. Schubiger (2013b) finds that communities exposed to direct and collective targeting by state forces during the Peruvian civil war were more likely to engage in counterinsurgent collective action at later stages. These findings are consistent with the notion that indiscriminate state violence may succeed in triggering civilian collaboration with the government if insurgents fail to provide protection (Kalyvas 2006), though pro- and counterinsurgent mobilization may occur in parallel, even at the very local level (Schubiger 2013a, 2013b). In other words, these findings do not indicate that indiscriminate state violence is “effective,” as it also tends to foster insurgent recruitment (e.g., Wood 2003). Militia mobilization may thus be one way through which cleavages are expressed at the local level, reflecting both prewar politics and processes endogenous to the war (Kalyvas 2006; Wood 2008).
Militias as Perpetrators of Civil War Violence
The reliance of state forces on militias may increase violence against civilians, as a result of both the state’s strategic incentives to outsource violence and the state’s lack of control over militias (Mitchell, Carey, and Butler 2014). The types of violence may be affected as well. Militias can support states in gathering high-quality local information and in forging the perception of state violence as selective, and they may also assist states in moving away from indiscriminate and toward selective violence (Kalyvas 2006; see also Lyall 2010 on the effectiveness of co-ethnic militias).
Besides the influence of the state on militia violence, militias’ internal characteristics have an effect on the level of violence. There is evidence that forced recruitment increases the likelihood of certain types of militia violence (Cohen and Nordås 2015), as violence is employed to strengthen in-group cohesion. Stanton (2015) finds that if militias recruit from the same social groups as the insurgents, they are less likely to target noncombatants, as this would imply attacking their own civilian constituency (Stanton 2015). Moreover, if social networks that underpin the screening and selection of new recruits fall apart over the course of the war, abusive behavior against civilians appears to become more likely (Forney 2015). Apart from these initial conditions of militia recruitment, recent research points to the relevance of armed group institutions in explaining patterns of insurgent violence (Gutiérrez Sanin 2008; Hoover Green 2011; Vargas Castillo 2011; Wood 2010, 2012). In fact, there is little reason to believe that militia combatants are less amenable to socialization and indoctrination than insurgent recruits.
Finally, the mobilization of militias is a key component in various counterinsurgency strategies. For a long time, the debate on counterinsurgency has centered on the dichotomy between coercive and public goods (“hearts and minds”) strategies. However, it is now apparent that the provision of public goods by the counterinsurgent is contingent on local security, which may be enhanced by militias. Furthermore, militias can be direct conduits of public goods provision too (Biddle 2008; Kalyvas 2008c).
The Transformation of Wartime Social Networks and Institutions
In contrast to most insurgent groups, counterinsurgent militias typically emerge not prior to, but during, civil wars, and as a reaction to wartime violence. In a study of militia formation in Greece during the 1940s, for example, Kalyvas (2008a) finds that the strongest predictors of militia recruitment at the village level were prior insurgent violence and incumbent territorial control. One hypothesis we derive from these reactive patterns of militia formation is that initial processes of mobilization are typically more localized than insurgent recruitment and more embedded into local institutions (see Forney 2015). Jentzsch (2014) for instance shows that in Mozambique, preexisting cultural practices provided the basis for the creation of new militia institutions.
The first implication for studies aiming to understand why some segments of the population engage in counterinsurgent collective action while others do not is that units of analysis must be granular. Second, scholars interested in the transformation of wartime social networks and local institutions should be attentive to the role militias play in the militarization of local governance and the polarization of public loyalties (Wood 2008; Schubiger 2013a). Although research on variations in wartime governance has made important inroads recently (Mampilly 2011; Arjona 2010; Weinstein 2006), the militarization of local governance through militias remains largely absent (but see, for instance, Fumerton 2001; García-Godos 2006; Schubiger 2013a, 2013b; Jentzsch 2014). This is surprising, especially given the long-term institutional and social legacies of militia activities in civil wars (Bateson 2013).
Recruitment
Five decades after Charles Tilly (1963) commented on the striking lack of attention to counterrevolutionary movements in studies of revolution, we still know surprisingly little about who initiates and joins militias and why. The scarce existing evidence on the determinants of individuals’ participation in insurgent and counterinsurgent armed groups shows that, first, social background variables are poor predictors, which runs contrary to what the extrapolation of classic grievance or opportunity cost arguments would suggest. Second, both pro- and counterinsurgent mobilization are driven by processes that often are endogenous to civil war (Humphreys and Weinstein 2008; Arjona and Kalyvas 2012; Kalyvas 2008b). In other words, the notion that pro- and counterinsurgent armed groups recruit from mutually exclusive social strata or different “types” of prospective recruits could be incomplete at best, misleading at worst. Instead, these studies suggest that preferences are heterogeneous across recruits within both “camps,” that multiple types of incentives may overlap at the individual level (Humphreys and Weinstein 2008), that endogenous dynamics are central (Arjona and Kalyvas 2012; see also Kalyvas 2008b), and that individual preferences may be transformed through the very experience of being part of a given armed group (Gutiérrez Sanin 2008).
Extending, building on, and challenging these insights, Oppenheim et al. (2015) study the determinants of insurgent defection in civil war. Specifically, the authors address the question of why some insurgents stay loyal to their armed group, while others do not, instead either demobilizing or defecting to counterinsurgent forces. Drawing on data from a representative survey of ex-combatants in Colombia, the authors find that the initial reasons for joining the insurgency condition the effects of wartime experiences, including indoctrination.
More generally, civil war research needs to take into account that rebel organizations compete for recruits with militias. We therefore advise caution in the interpretation of aggregate variables, such as unemployment rates or land inequality, that rely on the assumption that combatants are recruited primarily by rebel groups.
Implications for Theories on Post-Conflict Dynamics
The study of militias contributes not only to theories relevant for understanding conflict processes but more generally to theories of post-conflict dynamics, including postwar violence, inclusiveness, and citizenship. What are the post-conflict trajectories of militias? Are they repressed, integrated into formal structures of governance, or transformed into new armed groups, political or purely criminal? How does their action shape the post-conflict landscape?
The relationship between militias and the state poses important questions for state and nation building. The delegation of tasks to militias is puzzling insofar as they weaken Weberian states. Paradoxically, states have incentives to use militias as cheap instruments for the projection of state power, especially when the state security apparatus is weak (see Staniland 2015; Eck 2015; Carey, Colaresi, and Mitchell 2015). Civilian governments promote strong paramilitary organizations to counter any threat the regular military might pose to regime survivability as a form of clientelist politics (Dowdle 2007). More broadly, militias also have implications for the dynamics of nationalism, secessionism, and ethnicity. The successful use of militias to promote defection from an ethnic group may have long-term effects on ethnic identities, altering perceptions about their cohesiveness, and, therefore, the political viability of secessionist projects (Kalyvas 2008b).
However, the proliferation of armed groups may also undermine state authority. The use of militias can have unexpected and unwelcome consequences, including the violation of laws of war, the undermining of governmental authority, and the prospects of endemic internal strife (Hughes and Tripodi 2009). Militias may become instruments of specific elites used to further and perpetuate their domination (Schlichte 2010). Moreover, militias can have a considerable destabilizing effect on postwar societies and may be able to mobilize and reactivate social networks for illegal activities and influence electoral politics (Acemoglu, Robinson, and Santos 2009).
The effect of militias on postwar stability varies, however. Using subnational data from Colombia, Daly (2011) finds that the postwar remilitarization of paramilitaries depends on their wartime mobility. If members of a paramilitary unit come from diverse areas and are deployed far away from their homes, the group may likely disintegrate and lose its capacity for collective action after the war ends—a finding echoing past research on localized insurgent groups connecting locus of recruitment with military performance (Gould 1995). Rather than undermining or reinforcing the central state apparatus, militias can engage in forms of governance that parallel—and indeed, complement—state-enforced systems of order. For instance, in Guatemala, the civil patrols that emerged during the civil war turned into powerful enforcers of vigilante justice in the postwar period, effectively shielding communities from narcotrafficking cartels (Bateson 2013).
Given the tendency in the post–Cold War era of many civil wars to spawn democratization processes, we must understand the exact impact of militias for democratization and the quality of democracy. The militias-turned vigilantes may undermine the promotion of human rights and rule of law (see Smith 2013). In Colombia, paramilitaries are able to manipulate rural elites for electoral support and, therefore, skew the electoral game (López Hernández and Ávila Martínez 2010; Chaves and Robinson 2010). Militias (collectively but also individual members) may also become “re-tooled” for participation in criminal gangs and cartels. It is also likely that participation in militia activities gives rise to new political identities that may shape voting (and more generally political) behavior for a long time after the end of a civil war.
Conclusion: An Emerging Research Agenda
To summarize, the emergence of a new research agenda centered on the formation, activity, evolution, transformation, and devolution of various types of militias is important for both substantive and theoretical reasons. Substantively, it highlights an important dimension of civil war that hitherto has been neglected. Militias influence the length and outcome of civil wars, levels and types of violence, and contribute to the provision of public goods at the local level. The focus on militias forces us to revise several assumptions of theories that have omitted them. The study of militias also points to civilian agency in civil war, the ways in which macro cleavages are projected at the local level, and the transformative power of civil wars with respect to social networks and political institutions. Furthermore, theories of mobilization must incorporate the fact that various types of armed groups compete for recruits and civilian support.
The emerging research on militias must rely on clear concepts, disaggregate militia types depending on their function and evolution, pay close attention to variation in processes of militia formation and mobilization, and link theoretical and empirical findings to established theories of armed conflict as well as their gaps. It also needs to heed the call for distinguishing state-orchestrated and community-driven forms of militia formation and mobilization (Schubiger 2013a; Jentzsch 2014). Future work should continue to analyze militias with respect to their function and other dimensions of variation that promises to lead to better specified theories of militia activity in civil wars.
Potential research avenues for studies on militias also include the sources of variation in these dimensions (e.g., wartime local governance, violence against civilians, and armed group recruitment), the effects of state sponsorship on the behavior of militias, and the implications for theories of armed conflict and state building. Finally, research on internal armed conflict must depart from its tendency to equate the concepts of “armed group” and “rebel group” and abandon its canonical focus on state and insurgent actors as sole providers of wartime political order and violence.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.