Armed Group Structure and Violence in Civil Wars

The Organizational Dynamics of Civilian Killing

Roos Haer

• Publisher: Routledge
• Year: 2015
• Print ISBN: 9781138829367
• eBook ISBN: 9781317569176
• Edition: 1
• Topic: Politics & International Relations
• Subtopic: Military & Maritime History
• Index: Politics & International Relations

This book examines whether differences in the organizational structure of armed groups shape patterns of human rights violations in civil wars.

Since the end of World War II, civil wars have been characterized by extremely high numbers of civilian casualties. However, the exact extent of civilian suffering varies across time, conflict, and geographic region. Recently, a new strand of research has emerged, primarily focused on studying the dynamics underlying the variation in civilian abuse by examining the characteristics of the armed groups and how these characteristics influence the armed groups' behaviour towards the civilian population.

With reference to principal-agent theory and data on the organizational structure of more than 70 armed groups active worldwide from 1989 onwards, the author's analysis functions both on the level of the armed group and on the level of the individual via personal interviews with combatants.

Offering a unique insight into how factors such as recruitment methods, hierarchy and organizational commitment may affect the likelihood of civilian abuse by combatants, this book will be of much interest to students of political violence, civil wars, war and conflict studies, security studies and IR in general.

Part I

Theory

The internal organization of armed groups

1 Principal–agent theory and armed groups

Recent years have been characterized by a heightened academic interest in organizations. It is in the context of organizations that collective action is most effectively coordinated, that prisoner’s dilemmas are most readily overcome, and that stable social decisions are most likely made (Kiewiet and McCubbins 1991). The organizational bases of collective action are many – firms, bureaucracies, associations, committees, leagues, representative assemblies, but also armed groups. What the most prominent forms of organizations have in common, however, is the delegation of authority to take action from the individual or individuals to whom it originally endowed (Kiewiet and McCubbins 1991: 23). When talking about organizations and delegation, one quickly arrives at the agency theory that explains various aspects of delegation. In the following chapter, I explain the basic concepts of the principal–agent theory and how this is related to armed groups.

Development

Agency theory was first formulated in the field of economics. In 1971, Spence and Zeckhauser (1971) published an article in which the framework was presented that clarified the dilemmas of dealing with incomplete information in insurance industry contracts. By the 1980s, economists working on insurance models had defined the issues, concerns, and canonical result of the principal–agent theory (Holmström 1979; Shavell 1979; Miller 2005). From its roots in economics, agency theory has developed along two lines: positivist (empirical) and principal–agent (theoretical) (Jensen 1983). The two streams share a common unit of analysis: the contract between an actor that delegates and one that receives the extra responsibility. Both research strands also share common assumptions about people, organizations, and information. However, they differ in their mathematical rigor, dependent variable, and style. Positivist researchers have primarily focused on identifying situations in which the actors involved in the delegation are likely to have conflicting goals and they then describe the mechanisms that limit the actions of the actors that receive the extra responsibility (Eisenhardt 1989). Principal–agent theory researchers, on the other hand, are concerned with a general theory of the principal–agent relationship, a theory that can be applied to employer–employee, lawyer–client, buyer–supplier, and other agency relationships. This paradigm involves often careful specification of assumptions which are followed by logical deduction, and mathematical proof (Eisenhardt 1989).

Terry Moe (1984) made the translation of the agency theory to political science, where it has become very influential. However, within political science the differentiation between the two streams is less emphasized and one generally speaks about principal–agent theory without acknowledging the two strands. Most political scientists are, however, primarily concerned with the positivist stream of the agency theory. They apply the theory and describe the mechanisms that solve the problem of agency loss. Only few political scientists have been concerned with the theoretical principal–agent stream, i.e., which delegation contract is the most efficient under varying levels of outcome uncertainty, risk aversion, information, and other variables (Eisenhardt 1989). However, when political scientists first applied the principal–agent theory, not only were the two strands largely disregarded but also further differences emerged from the original economic formulation (Waterman and Meier 1998). In its new formulation, it was postulated that the principal and agents in political settings share the same characteristics as they would do in economic settings.

The first applications of this theory in political science was in the field of American and comparative politics, in which the theory has become the basis for an extensive set of studies relating to the delegation to bureaucracy and legislative committees (see e.g., Banks and Weingast 1992; Mitnick 1973; 1975; 1980; Moe 1982; 1984; 1985; Kiewiet and McCubbins 1991; McCubbins and Schwartz 1984; Wood and Waterman 1991; 1993; 1994; Scholz and Wei 1986). The most important finding from these applications is that the previously widespread view among scholars – that all-powerful bureaucrats often run amok in the policy process – is dramatically overstated, if not false (Nielson and Tierny 2003). The agency theory has also been taken up by international relations scholars. They have applied it to the study of international organizations, noting several similarities between domestic bureaucracies and multilateral bodies (e.g., Brown 2010; Pollack 1997; Nielson and Tierney 2003). Especially in the field of European Union (EU) studies, principal–agent models and the theory of delegation have been increasingly applied (Kassim and Menon 2003). This is in large part due to the fact that the theory is explicitly concerned with complex inter-institutional interactions, which is an inherit aspect of the EU (Kassim and Menon 2003). As a result, the principal–agent model offers a way of grasping the institutional complexity of the EU without relying too much on the intergovernmentalist versus neofunctionalist debate that historically dominated the field (Pollack 1997; Kassim and Menon 2003).

Principal–agent theory has achieved widespread recognition in political science and research has proliferated (see Miller 2005 for a good overview of the application of the principal–agent theory in political science). However, within security and conflict studies, principal–agent theory has seen relatively little use (Rauchhaus 2009: 873). Some scholars have, however, applied the theory to the study of civil–military relations. Deborah Avant (1993; 1994) was one of the first that borrowed insights from the principal–agent theory to explain different propensities for innovation across British and American military organizations. Risa Brooks (2000) also uses the theory to compare how different patterns of civil-military relations produce different grand military strategies. Amy Zegart (1999) uses it to explore the design of national security agencies at the start of the Cold War.

However, applications of the principal–agent theory to the study of armed groups and their behavior towards the civilian population almost do not exist (Johnson 2008). An important exception forms the study conducted by Gates (2002). Although only limitedly focused on the internal dynamics of armed forces, Gates (2002) developed a model to demonstrate that geography, ethnicity, and ideology play an important role in determining military success, deterring defection within the rebel group, and shaping recruitment. Also Hovil and Werker (2005) use the principal–agent theory as a basis in their study to anti-civilian violence in Western Uganda. They found out that a high level of violence towards the Ugandan civilian population was the result of a divergence of interests, unequal access of information, and contracting limitations between the financier and the insurgents. To put it more precisely, civilians were victimized by the Allied Democratic Forces (ADF), so that they could keep their outside funding. Also the study conducted by Butler et al. (2007) forms an exception and use the principal–agent theory to investigate sexual violence committed by government security forces. The authors start their theory with the assumption that when combatants are given a high degree of slack, they are more likely to perpetrate incidences of sexual violence. After controlling for factors likely to increase the incidence of sexual violence, such as military size and ethnic fractionalization, they found that, where agents are more accountable and subject to tighter control, sexual violence is less likely perpetrated. Note, however, that their dependent variable is the number of incidences of sexual violence, not the number of sexually violated victims.

Assumptions

The principal–agent model is an analytical expression of the agency relationship, in which one party considers to enter into a contractual agreement with another in the expectation that the latter will subsequently choose an action that produces outcomes desired by the former (Moe 1984). In other words, principal–agent scholars assume that social life is a series of contracts, in which the “contracts” are between (two or more) parties (Perrow 1986: 224). One, designated as the agent, acts for, on behalf of, or as representative for the other, designated the principal, in a particular domain of decision problems.¹ Examples of agency relationships are legion. For instance, in the doctor–patient relationship, the doctor is the agent hired by the patient, the principal. But the cooperative executive of a company and her subordinates as agents is also a good example. Note that the cooperative executive is in turn an agent for the shareholders (Pratt and Zeckhauser 1985: 2). In sum, essentially all contractual arrangements can be approached with this particular theory (Ross 1973).

As these examples tend to suggest, a principal may delegate to an agent for various reasons. Delegation is then also the basis of the principal–agent theory and several factors motivate a principal to give up some authority over outcomes by delegating it to an agent. One of the most important functions of delegation is to save costs. Delegation allows principals to internalize costs associated with arm’s length transactions as well as to realize efficiency gains associated with specialization. Principals may lack time and task-specific expertise to carry out all required operations and will consequently contract an agent to perform a specified job or jobs (Byman and Kreps 2010; Egan 1998; Salehyan 2010). They might also delegate to increase the credibility of their commitments or to displace responsibility for unpopular decisions (Fiorina 1977; Epstein and O’Halloran 1999). For example, delegations to enforcing agents with high discretion such as an independent central bank signals commitment, since granting enforcement authority to this agent makes it less possible for the principal to back out of its pledge (Byman and Kreps 2010). Also delegating warfare to private security companies is a good example of displacing responsibility for unpopular decisions, such as the use of violence against the civilian population. The principal might also seek to delegate to agents in order to ensure that the principal’s preferences are acted upon well beyond the duration of the principal’s tenure (Byman and Kreps 2010). Delegation resolves then also the problem of policy-making instability since delegation to an agent that shares and implements policies is one way to enact change over the longer term when the principal’s power may wane or when other principals may assume greater power (Byman and Kreps 2010; Pollack 1997; McKelvey 1976; Riker 1980).

Given some motivation for relying on an agent, the principal offers the agent a contract. The principal has the first mover’s advantage, which resembles the Stackelberg game in which the leader and the followers move sequentially (Von Stackelberg 1934). Consequently, the principal is endowed with all of the bargaining power and can make a “take-it-or-leave-it” offer to the agent (Sappington 1991: 47). However, the offered contract must satisfy the so-called participation constraint of the agent, i.e., the agent’s compensation must be at least as great as his opportunity costs. If this condition is not met, the agent will not be made better off by entering into the relationship and will decline to sign the contract. Assuming that the participation constraint is satisfied, the principal’s goal is to delegate tasks and responsibilities and to specify a corresponding schedule of compensation in such a way that the agent is motivated to best serve the principal’s interests rather than shirking or sabotaging (Kiewiet and McCubbins 1991: 27–28).

If information flowed costless and perfectly, principals would know what their agents knew and what they were doing. They could be confident that agents were operating as if they were principals themselves (Pratt and Zeckhauser 1985). But in real life, complete and full information is rarely freely available to all parties. Agents’ possess more information about their tasks than their principals do, though principals may know more about what they want accomplished (Pratt and Zeckhauser 1985: 3). It is the very nature of delegation that results in this information asymmetry. However, note that the exact nature of this information is unclear in most circumstances (Mitnick 1986). In the more mathematical applications of the principal–agent model, this is a lesser problem since ambiguities in these studies can relatively easy be removed. However, most principal–agent studies lack a clear definition or idea on what this information connotes (Waterman and Meier 1998).

In the study of armed groups and their behavior, much information is common to both the agent and the principal. For instan...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• List of figures
• List of tables
• List of abbreviations
• Acknowledgments
• Introduction
• Part I Theory – the internal organization of armed groups
• Part II The armed groups
• Part III The combatants
• Appendix 1: Organizational Commitment Questionnaires (OCQ)
• Appendix 2: Measurement of hierarchical structure
• Appendix 3: Web survey
• Appendix 4: Case selection
• Appendix 5: Combatant survey
• Appendix 6: Consent form interview
• Index