Dissertation Prospectus:
InsurgentStateBuilding
Ana M. Arjona
1. Introduction
The type of relations between insurgent and paramilitary armed groups, on the one hand, and the civilians whom they attempt to rule, on the other hand, is subject to wide variation. Some groups try to approximate the behavior of organized states by extracting taxes, imposing new social norms, establishing predictable and routinized systems of rule enforcement, and supplying public goods. For example, the FARC in Colombia has established, in many rural areas it controls, rules relating to a variety of behaviors—from criminal activities to fishing. In one particular region they even created an office of complaints open to the public[1] and distributed identity cards to civilians[2]. Other armed groups, however, behave like predatory bandits, reneging from ruling civilian affairs. The LRA in Uganda exemplifies this form of rule by continuously raiding local populations, stealing from them, and abducting close to 20,000 children to turn into combatants or sexual slaves. The range of variation is staggering in its range yet our systematic knowledge about it is close to nil, as is our theorizing about its causes and effects.In an attempt to make a step towards providing an understanding of these phenomena, my dissertation will address a set of critical questions: How do armed groups approach civilian rule, how do civilians respond to these approaches, and how does their response affect the behavior of armed groups? What explains the decision of civilians to collaborate with armed groups, enlist in their ranks, remain neutral, flee to other areas, or oppose them?
I aim to build a theory on armed groups’ ruling strategies and civilian behavior that delves into the organizational structure of armed groups; community structure; and individual characteristics. By theorizing on the interaction of these factors, I will assess the different mechanisms through which local dynamics change in the midst of war, and the ways in which this changes in turn shape armed groups’ and civilians’ choices. I will derive implications for specific choices civilians make such as enlisting in an armed group, providing voluntary support to it, and resisting its rule. By combining different methodologies and supplying new data on the Colombian case, collected to answer these questions in a systematic manner, I will be able to test my hypotheses in an innovative way.
Providing an answer to these questions on armed groups’ behavior is important for several reasons. First, the existing variation across these organizations in their approach to ruling civilians is puzzling: given that they all need to establish some sort of territorial control, why do they attempt to do so in such different ways? What are the strategies that ‘work’ and allow combatants to establish social order? What renders sympathies and voluntary support? What exacerbates hostility and breeds the emergence of new armed organizations seeking to resist? The importance of these questions lies not only on their relevance for understanding war dynamics, but also because they underlie more general questions about human behavior, political rule, leadership, and mobilization. Second, a theoretical understanding of armed groups’ strategies at the local level should illuminate questions on a variety of phenomena such as the intensity, form, and targeting of violence, the relation between local elites and the warring sides, and the possibilities of governmental or external intervention in ‘occupied’ territories.
How civilians react to armed groups’ attempts to rule them, the second research question of this project, is a relevant question for at least four reasons. First, it is an important enterprise in and of itself as it informs us about how individuals experience war. Second, a better understanding of civilian collaboration should also illuminate our approach to a variety of questions. If, as argued by many scholars and observers, civilian collaboration is the sine qua non of victory (e.g. Mao 1978), improving our knowledge of different types of collaboration should allow us to better assess questions such as the determinants of the warring sides’ capacity to expand and win, and war duration and termination. Third, a better assessment of civilians’ behavior should also illuminate the study of different attempts to build peace such as peace negotiations, demobilization programs, and the so-called peace communities because civilians’ choices can affect the outcomes of these initiatives. Fourth, unless we understand how civilians behave during war, we might not be able to identify and respond to the challenges involved in post-conflict reconstruction such as reinserting ex-combatants to civil life, recovering (or building) the rule of law within a community, controlling crime, and decreasing the odds of conflict resume. Finally, when researching both the dynamics and consequences of civil war, civilians have to be considered as agents; explaining the decision of civilians to comply with, or to offer different forms of support to, a particular armed group is a step forward in this direction.
Last, this project is worthwhile also because by investigating the processes through which populations’ allegiance is gained by groups attempting to rule them, the project will theorize of processes of state formation and social mobilization. In this sense, the collection of systematic empirical data that the project entails is not only important in its own right, but consequential insofar as it allows us to address in “real time”, so to speak, theoretical conjectures about the formation of states and the organization of polities that have so far relied primarily on historical guesswork or metaphors (e.g. Hobbes; Locke; Rousseau; Olson; Bates).
This prospectus is organized as follows: In section 2, I present a brief assessment of the existing literature on armed groups’ behavior towards civilians, and civilian behavior in the midst of war. In section 3 I frame the research questions that the project will address and discuss my dependent variables. In section 4 I present the independent variables and summarize the mechanisms through which they shape armed groups’ strategies and civilian collaboration. In section 5 I outline the research design.
2. Armed groups’ and civilians’ behavior in the literature
The political science literature on civil war has mostly focused on the following questions: what are the determinants of its onset? (Fearon and Laitin 2003, Humphreys 2003, Miguel et al. 2003); what explains variation in civil war duration and what are the determinants of its termination? (Fearon 2001, Elbadawi and Sambanis 2000); what determines the occurrence of specific kinds of crimes as genocide? (Harff 2003, Londregan and Poole 1990); what are its effects both at the national and international level? (Collier et al. 2001). While this literature has made a key contribution by identifying recurrent patterns across civil wars, it faces serious difficulties for showing what are the mechanisms that underlie those observed patterns. For example, while there is massive evidence of the correlation of GDP with civil war onset the theoretical understanding of this finding is not clear. Conducting research at a more disaggregated level of analysis, theorizing about local dynamics and micro-foundations, and using research designs that allow for testing the existence of correlations as well as the underlying mechanisms seems thus to be a necessary step towards improving our understanding of civil war.
In this line of reasoning, some authors have started to conduct very valuable work on the local dynamics of war shedding light onto a variety of topics such as the determinants of the use of violence against civilians (Gates 2002, Kalyvas 2005, Humphreys and Weinstein 2005), the success and failure of the warring sides in controlling the monopoly of violence (Gutierrez 2003), and the use of specific forms of victimization (Strauss 2004, Wood 2005). Although with their fine-grained analyses these works have made a substantial contribution by increasing our understanding of how wars are fought, the question of how civilians and combatants interact in their ongoing contact with each other has only been partially addressed. Anthropologists, political scientists, and historians have conducted invaluable work on case-studies that provide detailed accounts of the ways in which combatants and local populations interact with one another (e.g. Theidon 2004 on Peru; Ngoga 1998 on Uganda; Ferro and Uribe 2002 on Colombia; Benjamin and Demarest 1988 on Guatemala). These accounts, however, do not provide a general theoretical framework to understand variation in behaviors across groups and local communities.
An emergent literature on the organizational structure of armed groups has looked more systematically to the strategies these organizations employ towards civilians, using formal models (Gates 2002) and comparative analyses (Weinstein 2003). Comparative sociologists have also conducted research on the ways in which armed groups operate (Goodwin 2001). The former literature has made key contributions by addressing the implications that armed groups’ constraints and internal organization have on their choice of specific strategies like those related to recruitment. The latter has highlighted key aspects of armed organizations such as its resemblance with states. Yet, the dynamics involved in ongoing interaction between civilians and combatants, and their effects on the behavior of both, are not accounted for by neither.
Turning to civilian behavior, students of civil wars and rebellions have often stressed the importance of civilian collaboration for rebel groups’ survival and victory. However, collaboration itself is seldom explained. Even the term ‘collaboration’ is rarely unpacked and any action that favors an armed group is taken to be an ‘act of collaboration’ (a few exceptions are Petersen 2001, Wood 2003, and Kalyvas 2005). In terms of how the decision to collaborate with a group has been addressed, only a few authors have directly addressed the question. However, works on civil war onset, peasant rebellion, and recruitment implicitly or explicitly approach collaboration in one way or another. Three competing explanations dominate these literatures: civilians collaborate because of ideological reasons (e.g. Hobsbawn 159, Scott 1976), be there in the form of goals or principles; civilians collaborate because they fear reprisals; and civilians collaborate because it pays to do so, given the material benefits that are often involved (e.g. Lichbach 1994). More recent approaches highlight the importance of moral commitments and emotions (Wood 2003; Petersen 2001). Evidence suggests that traditional accounts cannot alone successfully explain civilian collaboration: in the case of ideology, if this explanation could account for all acts of civilian collaboration we should not observe cases such as the Segovias in Nicaragua where contra rebels gained the support of the peasants despite the Sandinista regime’s “ideological commitment to advancing the material well-being and political power of the country’s majority, the workers and campesinos.” (Schoreder, 2000:47). Fear of reprisals cannot account alone for collaboration since in many cases support entails spontaneous help to combatants in the absence of coercion that often involves high risks, such as the spontaneous provision of intelligence; Wood (2003) and Petersen (2001) offer numerous examples of this type of behaviors. Finally, many instances of collaboration cannot be explained on the basis of material benefits. Recruitment is illustrative: available evidence suggests that many civilians join armed groups even when they know they will not receive salaries or any access to spoils (e.g. Gutierrez 2004b). The preliminary results of a survey with Colombian ex-combatants suggest that about 7% of those who joined the FARC did so while working at coca plantations (where they get paid high salaries in comparison to most peasants) even though the FARC does not pay any salaries and most joiners know that when they enlist[3]. To be sure, the point being made here is not that available explanations cannot account for any instance of civilian collaboration: all of them seem to be part of the picture. The argument is, rather, that none of them can alone account for the variety of behaviors that count as collaboration. Hence, a more fine-grained analysis of what an act of collaboration may entail as well as of the motivations, opportunities, and beliefs that may underlie it is required.
3. Armed groups’ strategies and civilians’ behavior: framing the question
Different questions could be asked about armed groups’ behaviors towards civilians, including the following: What explains an armed group’s decision to attempt to rule one locality but not other? Why does ruling a specific locality prove to be of greater importance for one armed group than for others? What accounts for the variation across armed groups and localities in the ruling strategies that these organizations employ? Even though all these questions are to some extent relevant for comprehending armed groups’ behavior, the dissertation will deal only with the latter for two reasons. First, because the project aims to address the ways in which armed groups attempt to rule civilians and how civilians react to those attempts, rather than the ‘prior’ question of what makes some populations ‘worth’ conquering in the first place. Second, providing an explanation for the groups’ decision to expand in one direction and not in other requires taking into account considerations of military strategy, availability of economic resources in different geographical areas, state capacity to respond in different ecologies, and other factors that may make controlling a particular village or town more or less feasible and attractive. Building such a theory would require a separate research project.
Civilians’ response to armed groups’ ruling strategies may as well involve a large number of ‘observable’ conducts. Armed groups’ intervention in a particular locality may lead civilians to alter their behaviors vis-à-vis other locals, leisure, economic activities, political participation, and even religious practices. The dissertation will focus on the more narrow question of how civilians react to armed groups’ strategies in terms of their decision of whether to leave their hometown, stay and support the group, be neutral, or oppose it. However, in order to account for these choices the dissertation will delve into the mechanisms through which armed groups’ strategies shape local dynamics and individuals’ motivations and payoffs. Hence, my theoretical account of how different aspects of individual and social life are affected by the war should have implications for understanding behaviors different from collaboration and resistance.
What about the response of armed groups to civilians’ behavior? Once a group starts intervening in some way or another in a locality it may well decide what strategies to adopt in the future based on civilians’ early responses. If the goal of this project is to understand how armed groups approach civilians and how they respond, it has to account for the possibility that civilians are not just ‘receivers’ of armed groups’ moves but also act as ‘players’. As will be explained below, I will take this causal link into account by incorporating timing in the analysis in two ways. First, by inquiring about the existence of institutional learning: I will argue that a unified group that has been fighting for ten years can derive lessons about what works and what doesn’t. Hence, armed groups should be expected to be less capable of establishing their rule in their early attempts than later on. Second, armed groups may re-define some of their strategies towards a particular population depending on civilians’ reactions in the recent past. In consequence, I will address both actors’ behaviors such that civilian collaboration at time t2 is explained as the outcome of armed groups’ strategies at time t2, which are in turn a function of, among other things, civilians’ behavior in time t1.
In sum, the dissertation will focus on three research questions: First, what are the ruling strategies armed groups employ, and why do they choose them? Second, how do civilians respond to these strategies? And third, how do civilians’ responses in turn shape armed groups’ strategies? In what follows I discuss at greater length my two dependent variables: armed groups’ ruling strategies and civilian collaboration.
Armed groups’ ruling strategies
Students of civil wars and rebellions have often stressed the importance of civilian collaboration for rebel groups. As Kalyvas (2006) puts it, “it is widely argued that the outcome of the war hinges on the behavior of civilians” and “almost all writers converge in asserting that no insurgent movement can survive without ‘civilian support’ and neither can incumbent victory be achieved without it”. In his theory of violence in civil war Kalyvas (2006) reformulates this idea within a theoretical understanding of the dynamics involved in irregular warfare. According to him, civil wars are called ‘irregular’ if frontlines are absent. In such wars “the boundaries separating two (or more) sides in an irregular war are blurred, fluid, and porous” due to what he calls the ‘identification problem’—i.e. “the ability of irregular combatants to hide among the civilian population”. Given this absence of frontlines and the possibility of the ‘enemy’ hiding everywhere, armed groups need to achieve local control in order to sustain themselves, expand with the aid of civilians, and prevent the other side from conquering new territories. But how do these organizations attempt to achieve such local control?