Why do anti-rebel militias govern?
Organization, networks and paramilitary governance in Colombia
By Tim Stork
MA Thesis 17th of August 2015
Final version
12,200 words (excl. references)
Research Master Social Sciences
Lee J.M. Seymour
Supervisor
Conny Roggeband
Abstract
The study of armed group governance has centred most of its attention on insurgents, rather than on anti-rebel paramilitary and militia groups. Yet, paramilitaries vary greatly in their strategies vis-à-vis civilians in territories under their control. What explains this variance and why do paramilitaries govern? Crafted around an outlier case of militia governance from Colombia, this article utilizes a most-similar case study design to explain paramilitary governance. The findings suggest that paramilitaries perform governmental functions when state control over the militia’s ranks is low. Specifically, I illuminate three mechanisms linking low state control to a higher likelihood for militia governance emergence. First, paramilitaries autonomous from the state cannot draw on government funds and consequently depend on locals for resources and manpower – thereby attracting fewer ex-militaries to its ranks. Second, militias kept on a tight state leash are more likely to be drawn into the government’s counter-insurgency campaign, making up for offensive militias. Finally, loose state control makes governance worthwhile because government officials cannot flaunt the paramilitary’s public goods provisions as if performed by the state. As a result, rewards gained from governing are thus not shared with government officials. The study underscores the need for ‘bringing the state back into’ the analysis of non-state armed group governance.
Introduction
Paramilitary groups vary greatly in their strategies toward civilians in territories under their control (Gutiérrez Sanín & Barón, 2005; Arjona, 2014a). From 2001 until its demobilization in 2006, the paramilitary José Luis Zuluaga Front (Frente José Luis Zuluaga, FJLZ) operated as the de facto authority in the Colombian municipality of Sonsón. The FJLZ was led by commander ‘McGuiver’ and belonged to the Peasant Self Defense Forces of the Middle Magdalena (Autodefensas Campesinas del Magdalena Medio, ACMM), a paramilitary bloc (bloque) led by McGuiver’s father in law Ramón Isaza (‘El Viejo’) and consisting of six fronts among which the FJLZ. Unlike the other ACMM fronts (frentes), which largely abjured governing populations in the territories in which they operated, the FJLZ is an exemplary case of armed group governance. McGuiver wrote and implemented statutes, constructed roads, built a school, hospital and rehabilitation centre, and provided housing to the poor. In return, the FJLZ requested taxes from the population and ran protection rackets to finance its public goods provisions and sustain the
war effort against the insurgent Revolutionary Armed Forces of Colombia (Fuerzas Armadas
Revolucionarias de Colombia, FARC) and National Liberation Army (Ejercito de Liberación Nacional,
ELN). The level of governance evident in the case of the FJLZ is rare but not unique.
Mali’s strategy of preventing armed insurgency was based on outsourcing state functions such as policing and taxation to Touareg clans and manageable Arab factions and militias (Boukhars, 2013). In Iraq, Moqtada Sadr’s militia brigades fighting alongside Prime Minister Nouri al-Maliki’s security forces not only took up arms against the Islamic State, but also gave Koran lessons, undertook neighbourhood reconstruction, collected trash, and organized soccer teams through its explicitly non-violent branch Mumahidoon (Abbas, 2013; Katzman, 2014: 11). During Lebanon’s civil war, sectarian self-defense militias became state-like administrations providing social services despite brutality (Zahar, 2001; Kingston & Zahar, 2004: 90; Baylony, 2010: 47; 85-86). Anti-rebel peasant rounds (rondas campesinas) in Peru who protected communities against the Shining Path (Sendero Luminoso) insurgency, provided local security and had mechanisms for conflict resolution in place (Starn, 1999). Under the pretext of security and order, self-defense militias have used grave violence against drug gangs in Rio de Janeiro’s crime-ridden favelas, levying tax in return (Arias, 2006; Ribeiro et al, 2010). In Nigeria, the Oodua People’s Congress and Bakassi Boys ethnic militias protected locals against rife crime and ruled “instant” or so-called “juju justice” as well (Smith, 2004; Merz, 2010: 25).
What explains this variance in anti-rebel militia strategy and why do some paramilitaries choose to ‘govern’? Non-state armed groups often engage in some type of governance – that is, “the organization of civilian affairs within the territories where they are present” (Arjona, 2008). Most studies of armed group governance have focused their attention on insurgent governance (Wickham-Crowley, 1987; Kasfir, 2002; 2005; Weinstein, 2007; Mampilly, 2011; Arjona, 2014a), rather than the governing behaviour of anti-rebel paramilitaries and militias. However, like their insurgent counterparts, paramilitary strategies vis-à-vis civilians vary greatly.
This variance is particularly puzzling as paramilitaries are often believed to be nothing more than mere appendages of Colombian state elites (Richani, 2002: 52; Hristov, 2009: 59; Grajales, 2011), thereby assuming these so-called pro-government militias (PGMs) are intrinsically “subservient junior partners of government” (Staniland, 2015: 2). In line with this conventional wisdom, which dominantly conceptualizes state-militia relations as constituting supportive collaboration “with regimes straightforwardly outsourcing or delegating violence to militias” (Ibid.), it is widely argued that by cooperating with paramilitaries states gain plausible deniability for flagrant violence committed against civilians in the course of often brutal counter-insurgency strategies (Cohen, 2001: 108-109; Campbell & Brenner, 2002; Mitchell, 2004; Roessler, 2005; Ahram, 2011: 14; Reno, 2002; 2011; Acemoglu et al, 2010; Mitchell et al, 2014). Although the “plausible deniability” argument helps understanding violence devolution to non-state actors, it does not help understand why some paramilitaries choose to govern and equally are allowed to do so. The existence of elaborate patterns of paramilitary rule suggests that these assumed PGMs have moved beyond repressive violence alone (Gutiérrez Sanín & Barón, 2005; Arjona, 2014a).
Mancur Olson (1993: 568) argues that government for groups larger than tribes normally arises not because of social contracts or voluntary transactions, but because of rational self-interest among those who can organize the greatest capacity for violence. If Olson is correct, paramilitary governance in Colombia becomes even more puzzling as the government and its national army, not the paramilitary United Self Defense Forces of Colombia (Autodefensas Unidas de Colombia, AUC), have maintained the greatest capacity to organize violence throughout the conflict (Christia, 2012: 272). The two-part puzzle driving this research question is why a government that has the greatest capacity for violence would allow their paramilitary ‘partners’ to become potential local competitors in governance, and equally, why some armed groups enter into the laborious task of governing territories and populations that the government itself has often abandoned.
Existing explanations of insurgent governance perceive of governance as “the range of possibilities for organization, authority, and responsiveness created between guerrillas and civilians” (Kasfir, 2002: 4; Mampilly, 2011: 3-4) (Italics mine). Consequently, scholars have mostly explained armed group governance by specific armed group and local population characteristics (Wickham-Crowley, 1987; Kasfir, 2004; 2005; Weinstein, 2007; Arjona, 2009; 2014a; 2014b; Metelits, 2009; Mampilly, 2011), thereby obscuring the role of complex bargains, deals, and tacit understandings between governments
and non-state armed groups (Staniland, 2012) in partially determining the range of possibilities armed
organizations have for de facto rule. Overall the diversity of state-militia relations out there have been ignored by wrongly assuming anti-rebel paramilitaries are necessarily pro-government forces (Barter, 2013). Acknowledging the diversity of state-militia relations, implies that these should not be overlooked when explaining paramilitary governance and variance in paramilitary rule. Because whereas the state is largely absent from rebel and insurgent held territories, its presence is variable in zones held by anti-rebel militias.
This article provides a theoretical explanation for armed group governance in civil wars that takes into account paramilitary ties with government officials. Drawing on theory-building case studies from Colombia, the goal is to account for varying pathways to paramilitary governance evident in Colombia and possibly beyond. So far scholarly work has been largely deterministic in emphasizing how rebels able to draw from economic endowments (either natural resources or external support) will demonstrate far less concern for civilian welfare because they have a readily available financial pipeline (Weinstein, 2007). I develop an alternative argument that takes both supra-local imperatives and local motives seriously in explaining paramilitary governance, and accounts for variance in rule.
Specifically, I argue that the degree of state control over a paramilitary front determines the range of possibilities paramilitaries have to govern. Put simply, states have either low or high control over a militia force. The former are relatively autonomous militias while the latter tend to tap
from government networks for both resources and manpower. Autonomous paramilitaries have more incentives to govern, but are not always willing or able to do so in the face of local opposition. I identify three mechanisms that tie state control to militia governance, namely enlistment in government counter-insurgency campaigns, the recruitment of ex-militaries, and attribution of militia governance.
The article proceeds as follows. First, I assess existing theories of rebel governance and develop my own argument. Second, I present a research design able to understand armed group governance and evaluate existing causal claims. Thirdly, the empirical section discusses several cases from Colombia to reveal causal mechanisms and bolster causal identification. To assess the argument’s external validity and scope conditions, I briefly discuss other conflicts. The conclusion delves into the argument’s implications for theory and policy.
Theorizing paramilitary governance
A study on the determinants of paramilitary governance belongs to a growing body of research on intrastate warfare and the development of alternative, non-state political orders (Kalyvas et al, 2008; Staniland, 2012; 2014; 2015; Naseemullah, 2014; Naseemullah & Staniland, 2014). In contrast to the traditional Hobbesian view of violence as an expression of disorder (Hobbes, 1651 [2014]; Huntington, 1968), these studies argue “that order emerges when an actor successfully uses violence to control conflict” (Brewer Norman, 2012: 39).
The existence of armed group governance flows logically from the plausible, and oft-repeated claim that civil wars represent competitive state-building (Fischer & Schmelzle, 2009; Kalyvas, 2006; Marten, 2006; 2007; Menkhaus, 2006; Mukhopadhyay, 2009; 2014; Reno, 1998; 2002; 2011; Staniland, 2012; Wantchekon, 2004; Wood, 2003). As a result of competition, non-state armed groups fighting in civil wars often engage in some form of governance, ranging from the provision of public goods such as health care, education and a police force, to the performance of
regulatory functions such as the imposition of social norms and taxation schemes, and more symbolic practices such as the embrace of anthems, flags and currencies. In a time that “the condemnation
of non-state order as institutionally destructive has been replaced by its celebration as a vehicle of embedded forms of order and authority” (Meagher, 2012), commentators have argued in favour of hybrid security systems in which militias provide a potential basis for reinforcing frail statehood (Ahram, 2011; Menkhaus, 2006; Boege et al, 2008; Krause, 2012). A major characteristic of militias is their “conservative” agenda of guaranteeing security and protecting property (Jentzsch, 2014: 22). Hence, these non-state violence wielders do not promote a revolutionary agenda as they secure pre-existing rules (Ibid; Humphreys & Weinstein, 2008; J. Mazzei, 2009: 102; Starn, 1995: 565; Vlassenroot, 2009), though this of course varies as militias are more than mere functional auxiliaries and can pose a danger to the state. Hybridized violence might be a mean toward stability, but equally hollows out the state as the central government loses its grip on a militarized periphery (Long, 2008; Reno, 2011; Malejacq & Seymour, 2015). Then, why would governments allow some militias to turn into state competitors governing local communities?
Initially, the governing behaviour of non-state armed groups was conceptualized as a state-like contractual relationship between the ruler and ruled (Wickham-Crowley, 1987), mostly focusing on rebel and insurgent governance. Ever since Timothy Wickham-Crowley broke new ground, his research has been followed and largely supported by specific case studies of rebel rule in Côte d’Ivoire (Förster, 2010), Sri Lanka (Mampilly, 2007; 2011; Stokke, 2006), Uganda (Kasfir, 2002; 2005; 2008), Sudan (Mampilly, 2011), the Democratic Republic of Congo (Raeymakers et al, 2008), and Colombia (Arjona, 2009; 2014a). Broadly speaking, insurgent governance is defined on the basis of two characteristics (Arjona, 2008: 2-3). First, scholars note that governance emerges in areas where the armed actor has some territorial control – either partial or full; and second, governance entails the “establishment of rules or institutions to regulate civilian populations, the relation between combatants and civilians, or both” (Ibid.).
Prima facie, both defining characteristics of insurgent governance falter when confronted with
governing paramilitaries. Anti-rebel militias and paramilitaries are “armed groups that operate alongside security forces or work independently of the state to shield the local population from insurgents” (Jentzsch et al, 2015). Hence, state presence is much more variable in territories under paramilitary control than zones where insurgents rule – that is, after all, zones where governments are largely absent. Not only can paramilitaries share local governance with government officials or strike symbiotic relations (Acemoglu et al, 2010), there are indications that governments can influence the behaviour of militias and that the latter’s strategies are tied to government incentives (Stanton, 2015: 14). Simultaneously, militias have set up local forms of governance departing from existing state structures while currently operating with remarkable autonomy (García-Godos, 2006; Schubiger, 2013a; 2013b). In this article I argue that militias kept on a tight state leash, tend to be pulled into the government’s violent counter-insurgency and have less incentives to govern than groups operating under low state control. Before delving into the argument further, I now turn to conceptualizing paramilitary governance and discuss existing explanations for insurgent governance.
Paramilitary governance: concept, indicators, measures
Both the existence of insurgent governance and the form it takes vary widely across as well as within armed organizations (Bakonyi & Stuvøy, 2005; Arjona, 2008) – ranging from “elaborately patterned relationships” with civilians to “the absence of any patterned activity” (Kasfir, 2002: 4). After all, governance is “a system of rule that is as dependent on inter-subjective meanings as on formally sanctioned constitutions and charters” (Rosenau, 1992: 12).
So far the literature has sought to study variation in armed group governance along different dimensions – sometimes juxtaposing governance to violence, while at other times perceiving of violence and governance as not necessarily mutually exclusive (Arjona, 2008). Few authors have
conceptualized rebel governance dichotomously. Wickham-Crowley (1987) analysed whether
guerrilleros can become the legitimate ruling authority or not, while Claire Metelits (2009: 2) drew
on Olson’s often cited distinction between “stationary” and “roving bandits” to pit insurgents that provide public goods and thus govern, against those fighters that “forcibly extract resources without establishing a reciprocal relationship”.
In search of mechanisms explaining rebel behaviour and governance, scholars like Metelits cited work on state formation and rule by Olson (1993) and Charles Tilly (1990). For Olson so-called bandits have a choice of taking either a consensual or coercive approach in dealing with civilian populations, with some choosing to transit from roving to stationary banditry, not because of any moral impulse or social contract, but because the “monopolization of theft” and the protection of tax-generating subjects eliminates unproductive anarchy (Olson, 1993: 568). Following this Olsonian logic, paramilitaries are armed groups concerned with survival opting for more consensual accumulation strategies because the rewards are much greater than the mere looting roving paramilitaries depend on.
Tilly (1990) underlines that 17th century European rulers chose to govern because the hostile
external environment required the “organization of coercion” in order to foster a political entity’s “preparation for war”. In order to garnish sufficient funds to sustain standing armies, rulers started to prefer consensual taxation over more coercive methods since the latter increased the likelihood of resistance (Tilly, 1985). While consensual taxation allowed leaders to raise extensive resources, plan for the long term, and build a standing army, taxpayers benefited from increased order as well as the ability to voice their concerns to the leadership (Mampilly, 2011: 33). Thus, states make war and war makes states. This Tillyesque perspective has led many contemporary students of rebel governance to believe social contracts do not explain why rebels rule, but develop
as a result of rule over time. Olson and Tilly resonate throughout existing understandings of armed
Besides dichotomous conceptualizations of rebel governance, scholars have focused on one or more dimensions exploring variation along a continuum. For example, Mampilly (2011) studies the effectiveness of rebels’ “civilian administrations” along a continuum from low to moderate and high. Moreover, Kasfir (2005: 272; 2008: 3) focuses on three related dimensions, namely the extent to which rebels encourage civilian participation, set up forms of civilian administration, and organize civilians to generate high value goods and services.
Finally, analysts have proposed typologies of armed group governance based on variation along different dimensions. In Inside Rebellion (2007: 166), Jeremy Weinstein studies how democratic rebel governments are by analysing their performance of governmental functions along two dimensions, namely power sharing which varies from unilateral rule to joint civilian-military rule, and inclusiveness, which varies from non-participatory to participatory. More recently, Ana Arjona (2009; 2014a) proposed a typology of different types of wartime social order in which armed groups attempt to rule based on two dimensions: whether there is an underlying social contract between the rebels and civilians, and when such a contract is in place, whether the scope of the rebel intervention in civilian affairs is broad or narrow. According to Arjona, this variation yields three types of rebel rule, namely domination (when no social contract exists), surveillance or aliocracy (when a social contract exists and the intervention’s scope is narrow), and rebelocracy (when a social contract exists and the intervention’s scope is broad).
Like Arjona, I study paramilitary governance variation along the scope of its intervention in civilian affairs, and for analytical purposes – similar to Olson – distinguish between club and public goods provisions in accounting for variance in rule. Broad interventions in civilian affairs move beyond the provision of security through a police force and the imposition of regulatory functions like social norms and taxation schemes. Hence, these include the provision of public goods such as health care, education and infrastructure, and might expand to the embrace of symbolic practices (e.g. anthems, flags and currencies) as well. Conversely, militias that narrowly
intervene do not move beyond the provision of security as a public good, and possibly levy taxes and impose social norms. Mostly, their security provisions are of the club kind, only provided to a segment in society able and willing to pay taxes, or belonging to the same identity group.
Existing explanations for armed group governance
Then, how do scholars explain variance in armed group governance? Broadly speaking, existing explanations for insurgent governance fall within two categories. The first category hypothesizes the importance of pre-existing governance structures and population characteristics in determining rebel rule. Wickham-Crowley (1987) argues pre-war governance at the sub-national level matters because dissatisfaction with earlier rule shapes support bases. Equally, Kasfir (2004) notes that when rebel organizations devote efforts for mobilizing support before waging war, creating civilian administrations is easier.
For Mampilly (2011) pre-existing state penetration determines the effectiveness of rebel governance because those civilians living in a territory formerly deeply penetrated by the state, are more likely to demand rule congruent with their former experience. In similar vein, Arjona (2009; 2014b) argues that armed group strategies are dependent on rebels’ expectation of civilian resistance. She believes that in areas with strong local indigenous institutions present, a deep armed group intervention in civilian affairs is less likely as governance of these communities is more likely to spark resistance.
The second category focuses on armed group characteristics and war dynamics. Scholars who ascribe to the greed-thesis in the civil war literature (See Collier & Hoeffler, 2004) have argued that insurgencies able to draw on economic endowments (either natural resources or external support) will demonstrate far less concern for civilian welfare because they have a readily available financial pipeline (Weinstein, 2007). According to Weinstein’s rentier argument,
resource-poor insurgencies are more likely to develop complex governance structures to woo civilian support than resource-rich armed groups, as the former are dependent on taxation for continuation. Other than greed, Lee Seymour’s (2014) work suggests opportunism could explain armed group governance when there are incentives to do so, such as fleeing locals or being paid off by a state uninterested in governing. Similar to Weinstein (2007) and Olson (1993), Arjona (2009a) notes that armed groups can only reap the rewards of governance on the long run, and therefore rebels with short time horizons are less likely to govern. Moreover, Metelits (2009) believes that armed groups opt for coercion (against civilians) when other armed groups threaten their monopoly over resources – placing a premium on rivalry and territorial control in understanding micro-level conflict dynamics (Kalyvas, 2006; Arjona, 2009; Arjona & Kalyvas, 2008).
An alternative explanation for militia governance
Whereas existing explanations for rebel governance note that the interaction between insurgents and local populations determine the range of possibilities armed groups have to govern, I argue – that in understanding paramilitary or militia governance – relations between state actors and paramilitaries should not be overlooked. After all, in contrast to territories where rebels govern and the state is largely absent, relations between government representatives and militia leaders are variable in zones under anti-rebel militia control. By definition, paramilitaries are groups that operate outside the regular command structure of government military and police forces. Consequently, scholars have argued that governments may have difficulty controlling militias, leading to higher levels of violence toward civilians (Mitchell et al, 2014). Kalyvas (1999) notes militia involvement in civil wars can lead to violence escalation – though, his argument stresses the role local-level cycles of revenge and retaliation have in explaining escalation, rather than government incentives to outsource violence.
Governments can reap the rewards of a brutal counter-insurgency strategy by cooperating with militias while simultaneously denying responsibility (Campbell, 2000; Campbell & Brenner, 2002; Ahram, 2011; Mitchell et al, 2014; Carey et al, 2015). Though the “plausible deniability” argument provides a framework for understanding violence devolution, it proves to be of little help to understand why some militias govern. Moreover, if the plausible deniability argument is correct one would expect to see regular government forces refrain from civilian targeting to maintain the government’s ability to deny responsibility for abuses, while – simultaneously – militias engage in higher levels of violence against civilians. Yet, militia forces rarely use higher levels of civilian targeting than their government counterparts (Stanton, 2015: 19). Instead, “when the government’s regular military forces target civilians, militia forces are also more likely to target civilians.” Jessica Stanton’s findings support the hypothesis that there is strong correspondence between the behaviour of governments and militia collaborators. Her findings suggest that militias receiving government training are more likely to engage in violence against civilians, bolstering the claim that militia strategy is not simply a by-product of the local conflict context, but to an important degree is tied to government incentives as well (Ibid: 14).
Accordingly, I argue that the degree of state control over the militia force determines both the opportunities and incentives paramilitaries have for governance. The degree of control state agents exercise over paramilitaries is conceptualized as either ‘low’ or ‘high’. Control determines whether non-state coercion is deployed for purposes congruent to state interests or not (Malejacq & Seymour, 2015: 2). Whereas most studies focus on the implications of state control for the use of violence against civilians by militias (Dasgupta, 2003), and often argue that states need to reign in militias to minimize the chance of predation and civilian abuse, I find instead that militia autonomy from the government can itself open up opportunities and incentives for the performance of governmental functions by paramilitaries. Particularly, I identify three mechanisms linking state control to militia governance: 1) participation in government
counter-insurgency (COIN) campaigns, 2) recruitment of ex-military personnel, and 3) militia attribution of public goods.
To measure state control over a militia group, the empirical section delves into the command structure linking government and militia forces, and examines the organizational ties connecting paramilitary commanders to government military forces, police officers and local political elites such as mayors, senators and governors. Indicators of state-militia collusion are armaments, shared patrols, paramilitaries receiving government training, and corruption charges or indictments linking politicians to paramilitaries. Yet, collusion does not necessarily imply control. Rather, state control results from different configurations of elite networks and relative power (Malejacq & Seymour, 2015).
Networks are “voluntary, reciprocal, and horizontal patterns of communication and exchange” (Keck & Sikkink, 1998: 8). Often they are based on flexible links between component organizations pursuing mutual interests (Powell & Smith Doerr, 1994: 323). If groups need to cultivate trust in conditions where hierarchical links prove disadvantageous, such as in civil wars, “networks offer an effective alternative form of organization that can facilitate political activity in circumstances where collaborators lack […] formal protection” (Arias, 2006: 7). Especially, networks allow groups to collaborate with a set of multiple specialised actors – while sharing needed information, and withholding confidential information (Ibid; Moss Kanter, 1991: 63-87). Mutual observation of member groups helps building organizational trust, transmits norms, and transfers legitimacy (Granovetter, 1983: 212-213; Uzzi, 1996: 674-678; Podolny & Page, 1998: 57-76; Diani, 2000: 291; Putnam, 2000).
State control over a militia is high when elite networks of paramilitary leaders and government officials overlap and the government is relatively more powerful locally. Local relative power results from military base presence and government troop deployments in the area. Overlapping elite networks harbour trust relationships from which both the ruling coalition and militia draw
their support and resources. When these networks do not overlap but are discrete, and the militia operates relatively isolated from a military base, state control is low – implying non-state coercion is more likely to be deployed for purposes incongruent with state interests, underscoring the militia’s autonomy.
Figure 1 provides a graphical summary of the theoretical argument, which identifies three mechanisms effectuating from state control linking higher control to a lower likelihood of paramilitary governance and vice versa. First, paramilitaries operating under high state control are more likely to be drawn into the government’s counter-insurgency strategy, and therefore tend to be mostly offensive – searching for enemies across territory. As Olson (1993) suggests, mobile bandits tend to be more abusive than stationary ones and have little interest in the prosperity of a given community. Conversely, loosely controlled paramilitaries or militias operating under no control whatsoever cannot tap from government support networks and thus are (more) dependent on locals for both resources and recruitment, underscoring they do have incentives to govern.
Figure 1. Causal relationship between state control and militia governance
Second, tightly controlled paramilitaries tend to have relatively more ex-military personnel within their ranks, while loosely controlled militias tend to attract more locals. Increased local presence within the militia force allows for ruling practices to be endogenously influenced and social contracts to be struck more easily. Part of the logic spelled out here comes from Weinstein (2007), who notes that exogenous resource availability, or better put in this case – the recruitment of ex-militaries – translates into predatory groups less inclined to be concerned with civilian welfare. Mediated by state control, recruitment’s invisible hand enforces different armed group strategies vis-à-vis civilians.
Third, autonomous militias are more likely to govern because governance is worthwhile when ties to the state are minimized. After all, under these conditions, corrupt, cost-conscious state officials cannot flaunt the paramilitary’s public goods provisions as if undertaken by the state. Hence, governing action undertaken by militias will not be attributed to government elites but to the paramilitaries instead. As a result, the rewards of governance are not shared with state actors, implying its gains, such as civilian collaboration and support are maximized in paramilitary hands
when state control is low, an important gain because civilian collaboration is widely acclaimed to be the sine qua non of victory in civil wars (Trinquier, 1964: 29; Mao, 1978; Guevara, 1997; Wood, 2003; Kalyvas, 2006; Arjona, 2009). Ultimately, not being kept on a tight state leash implies militias can contest the rulings of government officials and thereby situate themselves as the sole guarantors of order and protection. By avoiding the state or keeping its representatives at bay, paramilitary governance becomes a vector to enhance the militia’s local legitimacy vis-à-vis an often corrupt government alternative. Consequently, these militias are better suited to cooperate with local civil society organizations such as juntas de acción comunal and even labour unions, strengthening the militia’s ability to monitor communities.
Although the argument underscores the importance of state control in explaining militia governance, its understanding of paramilitary motivation is not state-centric. Most studies of paramilitary behaviour have ascribed to a state-centric understanding of paramilitary motivation
(Mason & Krane, 1989; Cohen, 2001: 108-109; Campbell & Brenner, 2002; Mitchell, 2004; Roessler, 2005; J. Mazzei, 2009; Ahram, 2011: 14; Reno, 2002; 2011; Acemoglu et al, 2010; Mitchell et al, 2014), overlooking the fact that paramilitaries are actors of their own and wrongly assuming that being anti-rebel logically implies being pro-government too (Barter, 2013). Conversely, I argue militia agency can be predicated on willingly avoiding government ties – implying state control is just as much a product of interaction, as a result of on-the-ground conditions and capacity.
Research design
The ongoing Colombian conflict can be traced back to a civil war called La Violencia, taking place between 1949 and the early 1960s. During that period, fighting between members of the two main political parties, the Conservatives and Liberals, caused the death of around 200,000 people (Roldán, 2002: 299-300; Arjona & Kalyvas, 2008: 14). After over half a century of bi-partisan
antagonism,1 the two parties established a rotating coalitional rule that is now better known as the
National Front of 1958 (Hartlyn, 1988: 26; Murillo, 2004: 45). With this shift towards power-sharing consociationalism an elitist hold on Colombian politics that remains until today, made-up for an “elite-friendly” democracy in which non-state armed groups proved pivotal (Robinson, 2013: 44).
Even though violence decreased following the National Front agreement, leftist guerrilla groups such as the FARC and ELN emerged in the sixties as the coalition’s promises for land reform were not kept.2 In the mid-seventies, guerrilla groups began to expand into new areas due to
growing revenues from illicit drugs, kidnapping and extortion (Centro Nacional de Memoria Histórica, 2014). Clearly, these methods affected the livelihoods of regional elites, such as land owners (latifundistas), cattle-ranchers (ganaderos), emerald traders, and political functionaries (Arjona & Kalyvas, 2008: 14). In response to a rising, often extortionist guerrilla, elites began to form paramilitary groups. Although some paramilitaries were truly self-defense groups formed by peasants, the majority was set-up by these powerful local and regional elites, along with emerging drug lords (Ibid; Duncan, 2007; J. Mazzei, 2009; Molano, 2009; Ronderos, 2014).
Why Colombia?
Studying the emergence of paramilitary governance in Colombia provides a unique opportunity to evaluate the complex mechanisms through which militias rule. The Colombian case is advantageous from a theory-building perspective as it illustrates significant within-case variation. In Colombia, paramilitaries have partly emerged in response to rebel violence (J. Mazzei, 2009). As a result, they had strong incentives to construct orders, privileging local security and
1 La Guerra de Mil Dias, a war that took place between 1899-1901, is often seen as “the first major clash
between representatives of the Liberal and Conservative parties” (Ruiz, 2001: 40). The three-year war accounted for 100,000 deaths.
2 Besides the FARC and ELN, the Popular Liberation Army (Ejército Popular de Liberación, EPL), the 19th of
April Movement (Movimiento 19 de Abril, M-19), and Quintín Lame emerged. The latter two are demobilized and no longer active.
governance over national mobility and organization (Gutiérrez Sanín & Barón, 2005; Gutiérrez Sanín, 2008; Álvaro, 2009; Brewer Norman, 2012; Barter, 2013). Besides the emergence of local paramilitary orders in which paramilitaries remained largely stationary and became the de facto authority, there is evidence of paramilitaries operating as mobile death squads in pursuit of the guerrilla, often committing grave violations of human rights (HRW, 1996; 2014; ICG, 2007; J. Mazzei, 2009). Other paramilitaries have been armed and trained by state officials as part of a government counter-insurgency strategy, or colluded with local and regional politicians and military commanders to deter guerrilla and share in nepotistic gains – giving rise to allegations of a parapolítica, or so-called “para-politics” to be in place (HRW, 2000; Ramírez, 2005; Álvaro, 2007; 2009; Barrera & Nieto Matiz, 2010; Acemoglu et al, 2010; Escobar, 2013). The nexuses that tie paramilitaries to government elites imply it has been unclear at times to whose interests these militias were aligned (Palacio Castanâneda, 1991; Gutiérrez Sanín & Jaramillo, 2004). After all, non-state armed actors in post-Cold War Colombia have predominantly self-financed (Guáqueta, 2003) – placing a premium on local dynamics, alliances, motivations and support as well.3 Finally,
the demobilization of the AUC and its separate blocs in 2006, among which those groups studied here, has increased data availability as subsequent legal processes opened up novel data sources and increased knowledge of these groups’ past behaviour. The research design presented here, purposively selects cases that are largely constant along all theoretically relevant independent variables, but vary dramatically at the dependent variable paramilitary governance.
Case selection
Specifically, the study presents a most-similar design crafted around an outlier case, the José Luis Zuluaga Front (FJLZ) – a paramilitary front belonging to the Peasant Self-Defense Forces of the Magdalena Medio (ACMM). In 2000, the ACMM’s commander Ramón Isaza decentralized his
3 See Medina Gallego (2005) and Romero (2011) for descriptions of the Colombian paramilitary’s finance
bloc along family lines by splitting up the ACMM in six fronts – with Isaza himself leading Frente
Central (FC). His sons, Oliveiro and Ovidio Isaza, alias ‘Terror’ and ‘Roque’, became the leaders
of Frente Heroes del Prodigio (FHdP) and Frente Jhon Isaza (FJI) respectively. Ramón’s son in law, Luis Eduardo Zuluaga – better known by his alias McGuiver – was appointed command of the FJLZ, while Isaza’s adoptive son, Walter Ochoa Guisao (‘El Gurre’) led Frente Omar Isaza (FOI). The sixth front was led by John Fredy Gallo Bedoya, ‘Pájaro’, who had no known family ties to Isaza and took control of Frente Celestino Mantilla (FCM).4
Commander McGuiver’s FJLZ was the only ACMM-front that wrote statutes, built roads, an electricity plant, houses for the poor, health care and rehabilitation centers, schools, and recreational facilities. The other five ACMM-fronts did not provide public goods or did so to a lesser degree (See Table 1). Hence, ACMM governance strategies varied in spite of similar conflict histories and the fact that the ACMM’s command was largely organized along family lines.
Table 1. Variation in public good provisions ACMM5 ACMM Front
Public good provision and bureaucracy Frente Central Frente Celestino Mantilla Frente Heroes del Prodigio Frente Jhon Isaza Frente José Luis Zuluaga Frente Omar Isaza
Written statutes No No No No Yes No
Roads (in km) 0 0 0 0 176 0
Electricity plant No No No No Yes No
Houses for poor
people Yes/No No No No Yes No
Health centers No No No No Yes No
4 Audiencia versión libre postulado Ramón Isaza Arango ante Justicia y Paz, 8th of June 2007; and, Exposición
sobre los orígenes y estructura de las Autodefensas Campesinas del Magdalena Medio, Fiscalía 2 de Justicia y Paz, audiencia de control de legalidad postulado Ramón Isaza Arango, 26th of October 2011.
5 Adopted from Bautista et al, 2013. Three interviews (#3, #4, #5) held with legal investigators researching the
crimes of the ACMM, suggest that apart from the FJLZ, all fronts scored negative on public goods provisions, calling into question the coding of Bautista et al. Therefore ‘Yes/No’ in Frente Central.
Schools No No No No Yes No Youth and old-age
rehabilitation center Yes/No No No No Yes No Sports and recreational facilities No No No No Yes No
Studying strategic variation in one bloc is particularly interesting because it allows one to keep salient variables in existing explanations for armed group governance, such as certain armed group and population characteristics, largely constant. Intuitively, case selection has been influenced by J.S. Mill’s (1872) “method of difference”. Mill’s method is particularly relevant to the study presented here because of its applicability to small-N data (Skocpol, 1986; Seawright & Gerring, 2008). Numerous influential small-N studies in political science applied Mill’s logics to infer causality (Moore, 1966; Skocpol, 1979; Katznelson, 1985; Stepan, 1985). Still, the method of difference makes a variety of assumptions that are unrealistic for social processes and could lead to serious distortions if not addressed (Nichols, 1986; Lieberson, 1991: 312).6 Moreover, the
conceptual “ladder of abstraction” should not be used too freely because then the similarities could outnumber the observed differences thereby biasing results further (Sartori, 1970).
Map 1. Geographical presence of ACMM fronts and ACPB bloc7
6 Mill’s methods assume a deterministic approach over a probabilistic one, no errors in measurement,
mono-causality, and the absence of interaction effects.
7 Adopted from Tribunal Superior de Bogotá, Sala de Justicia y Paz, Sentencia de Ramón Isaza Arango y otros,
chapter ‘Contexto Histórico Y Socio Político: Autodefensas Campesinas Del Magdalena Medio’. Published 29th of May 2014. Presence of ACPB in Puerto Boyacá was added by myself.
To address the weaknesses inherent to Mill’s methodology, the most similar design triangulates its findings from the comparisons of the six ACMM-fronts’ strategies by analysing the ruling behaviour of another paramilitary bloc operating in the Middle Magdalena as well, namely the Peasant Self Defense Forces of Puerto Boyacá (Autodefensas Campesinas de Puerto Boyacá, ACPB). The Middle Magdalena region spans five Colombian departments (Antioquia, Boyacá, Bolívar, Cesar, and Santander) and dozens of municipalities. Moreover, the ACMM operated in Tolima, Caldas and Cundinamarca as well (See Map 1). Of course conditions varied across all these sub-regions. For example, the nature of paramilitarismo in Víctor Carranza’s emerald region in eastern Boyacá differed drastically from the paramilitary activities in the coca cultivating territory of south Bolívar (CINEP, 2003; Gutiérrez & Jaramillo, 2004; Vargas, 2009; Brewer Norman, 2012:
49-50), or elsewhere.
The ACPB provides an interesting case as it has consistently governed its territory in close proximity to the ACMM, whose FC and FHdP operated in Puerto Triunfo and Puerto Nare, just at the other side of the Magdalena river opposite to Puerto Boyacá (Boyacá) where the ACPB ruled (See Map 1). Besides geographical proximity, the two blocs were formally united from 1984
until 1993, making variance in rule during this period, when the ACPB was led by Henry Pérez, and later by his successor Arnubio Triana Mahecha, alias ‘Botalón’, even more interesting. Like the ACMM’s FJLZ, the ACPB has provided public goods and undertook social welfare programs.
This article is based on three months of fieldwork carried out in Colombia and builds on seven data sources: interviews with experts, victims, ex-combatants and local leaders; public hearings of demobilized paramilitary commanders and combatants; video material from these interrogations; fieldwork reports of legal investigators; numerous reports and analyses from VerdadAbierta.com; ACMM and FJLZ-written documents; and the secondary literature on the regional and Colombian conflict.
Theory-building case-studies
The theoretical section has identified three mechanisms linking low state control to paramilitary governance. In the next section, I present evidence from the selected Colombian cases explaining how these mechanisms shape incentives to govern – underscoring the determinative power of state control, meaning state control over militia groups decreases the likelihood of paramilitary governance to emerge. Yet, before each individual mechanism is discussed, evidence is provided illustrating variance in both governance and state control among these cases.
Preliminarily, variance in governance within the ACMM has been illustrated, showing variation in public goods provisions among the six fronts (Table 1). In contrast to the other ACMM fronts, McGuiver’s FJLZ provides an exemplary case of armed group governance. After paying and provisioning for the FJLZ’s military organization, which accounted for 34 percent of its expenditures, a total of 23 percent was allegedly spent on public goods or what McGuiver called
“social works” (obras sociales).8 The FJLZ operated in the municipality of Sonsón and was the de
facto authority in La Danta, La Unión and Aquitanía, all located in the department of Antioquia. The FJLZ wrote and implemented statutes, constructed roads, built schools, houses for the poor, a hospital and rehabilitation centre, bull ring and football pitch. Also, the FJLZ organized multiple events, among which the hosting of a famous Colombian TV show in La Danta – an experience many civilians still remember vividly (Gallego Castro, 2013: 54). In return, the FJLZ requested taxes from the population and ran protection rackets to finance its public goods provisions and warring efforts against the guerrilla. In order to collect these taxes, which McGuiver called contribuciones or contributions instead,9 the commander had ‘civilian’ tax
collectors on the payroll.
These so-called financieros were explicitly separated from the FJLZ’s armed wing, an organizational practice that was common among other AUC-paramilitaries in the Middle Magdalena as well. For example, a social worker in Barrancabermeja remembers “that for every armed paraco [paramilitary] there were five unarmed paramilitaries operating alongside”.10 Others suggest
paramilitaries outsourced non-violent tasks to taxi drivers, primarily moto taxis, who then picked up extortionist vacunas and shared information – though these collections were often arbitrary and lacked any clear pattern.11 Tax collections followed according to well-defined rules in McGuiver’s
zone of control and taxes levied remained relatively constant over time, although the amount paid differed from farm to farm.12 In fact, that taxes remained constant over time suggests
McGuiver did not provide public goods to maximize fiscal revenues. After all, taxes did not increase as public goods provisions expanded. Illustratively, increased public mobility as a result of built road networks surrounding La Danta, did not lead to tax increases.
8 Accounting report FJLZ. Copy of report in possession of author.
9 See: Estatutos de la organización Frente José Luis Zuluaga, p. 14. Copy FJLZ’s statutes in possession of
author.
10 Interview #1, 16th of December 2014, Barrancabermeja. 11 Interview #2 15th of December 2014, Barrancabermeja.
12 Accounting report FJLZ; and, Interview #3, 21st of January 2015, Bogotá; and, Interview #4, 21st of January
Though the FJLZ provided public goods and wrote statutes intended to manage relations between paramilitaries and civilians, this does not imply McGuiver’s men were never abusive. Moreover, the norms FJLZ paramilitaries imposed were authoritarian to say the least. For example, the FJLZ’s statutes explicitly note that civilians transgressing the organization’s rules can be expelled from the region.13 Also, under the banners of a limpieza social (“social cleansing”)
the FJLZ held mandatory HIV check-ups for prostitutes (Gallego Castro, 2013: 54). Besides expulsion and displacement, official sanctions included forced labour and degradation.14 The
former included building roads, while the latter were for example “exemplary punishments” in which paramilitaries urinated over vagabundo teenagers, or cut off their long hair using a machete.15 Like McGuiver, Ramón Isaza imposed social norms, for example punishing
transgressors by forcing them to work on a yucca plantation (Verdad Abierta, 2014). Yet, Isaza did not choose to provide public goods, showing one could maintain territorial control without going that extra mile.
Then, why did McGuiver govern while the others simply ruled by cultivating fear alone? Primarily, I argue, the FJLZ’s governing strategy is grounded in McGuiver’s ability to avoid political elites and the Colombian military. As a result, state control over the paramilitary faction was low. The other front commanders, however, colluded with local politicians, shared in nepotistic gains or went on patrols with military battalions, thereby – as will be discussed in detail – minimizing the likelihood armed group governance would be both a possible and worthwhile strategy.
Variance in state control
13 Estatutos de la organización Frente José Luis Zuluaga, p. 19. 14 Estatutos de la organización Frente José Luis Zuluaga, p. 10.
15 Interview #6, 25th of November 2014, Bogotá; and, Interview #7, 25th of November 2014, Bogotá; and,
Undoubtedly, the fact that La Danta and other areas under FJLZ control are relatively peripheral zones beyond the military’s reach, is one reason McGuiver was able to keep military and state officials at bay. On the contrary, Ramón Isaza, who led Frente Central and remained the ACMM’s general commander after 2000, had maintained relations with government actors ever since he first mobilized in the late seventies. Isaza, alias ‘El Viejo’, was born in the municipality of Sonsón and moved to Las Mercedes (Puerto Triunfo) during his childhood (Gallego Castro, 2013: 45). It was there, along the banks of the Magdalena river, that he and seven other campesinos formed a self-defense group called Los Escopeteros in 1977.
These “Shotgunners” united in response to guerrilla expansion in the area – and specifically, against the extortion and kidnapping of cattle ranchers by the FARC’s 9th Front.16 Isaza had
worked as a local guide with the Colombian military the months before17 and even spoke to a
general in September 1977, seeking protection against arbitrary quotas and cattle rustling perpetrated by the insurgency. Allegedly, as Isaza recalls, the general said the army lacked manpower to patrol the area, meaning – according to El Viejo – “we had to defend ourselves as good as we could”.18 With the military unable to provide manpower, Isaza solicited the help of
local cattle ranchers and loggers, who gathered one million Colombian pesos among each other to form Los Escopeteros.19 Though the military was unable to patrol Las Mercedes, it did arm
Isaza’s men with its first eight escopetas, 12 mm calibre shotguns – most probably paid with the money gathered by the cattle ranchers, loggers and local strongmen.20
16 Audiencia versión libre postulado Ramón Isaza Arango, Justicia y Paz, 30th of April 2007, morning session. 17 Audiencia versión libre postulado Ramón Isaza Arango, Justicia y Paz, 24th of August 2007, morning session. 18 Audiencia de legalización de cargos de Ramón Isaza Arango y otros, Justicia y Paz, exposición de Fiscalía 2,
26th of October 2011. 19 Ibid.
20 Exposición del fiscal ante la Sala de Justicia y Paz del Tribunal Superior de Bogotá de Justicia y Paz, en
Between 1980 and 1984, Los Escopeteros became more offensive and known as “Los Masetos” – a colloquial reference to the paramilitary Death to the Kidnappers (Muerte A Secuestradores, MAS),21
a group co-found and -financed by 223 Colombian capos and multiple military officers. Quickly, MAS evolved into a “bridge” connecting the drug cartels and Colombian military in their shared struggle against the leftists (Dale & Marshall, 1998: 89; 94; Resselaer, 1991; Ronderos, 2014: 42). At that time, “Los Masetos and the military frequently patrolled the zone together. Along the highway they operated shared road blocks, while in towns and municipal capitals they threatened peasants to pick sides. It was assumed that whoever was not with them, was against them.”22
By the early 2000s, when Isaza decentralized his command over the ACMM, the commander’s involvement in government networks had only increased. During the paramilitary bloc’s twenty-third anniversary, local government officials such as Justo Capera Caicedo, the mayor of La Dorada, were present. Capera Caicedo gave an honorary speech during the ceremony and congratulated Isaza with “awards” (El Tiempo, 2008; 2009). Former mayor of Puerto Salgar, Villers Lozano, ex-congressman Carlos Clavijo, and another ex-mayor from La Dorada, Orlando Echeverry, joined the celebration as well, video recordings suggest (Ibid.). Recently, the Justice and Peace Tribunal denied to liberate Isaza because it is believed he has not told the complete story about his involvement in para-politics (El Tiempo, 2014).
Isaza’s involvement in government networks, starkly contrasts McGuiver’s purposive policy to avoid politics.23 However, like Isaza, his adoptive son El Gurre, the commander of the ACMM’s
FOI colluded with powerful politicians. For example, Javier Ramiro Devia Arias, Tolima’s departmental representative in Colombia’s Congress was convicted for his ties to FOI. The politician paid the paramilitary protection money during his electoral campaign in 2002. “During
21 Tribunal Superior de Bogotá, Sala de Justicia y Paz, Sentencia de Ramón Isaza Arango y otros, chapter
‘Contexto Histórico Y Socio Político: Autodefensas Campesinas Del Magdalena Medio’, p. 237. Published 29th of May 2014.
22 A testimony from a campesino civilian living in a zone where Los Masetos operated, recorded on the 9th of
March 1989. See: Grupo de Memoria Histórica, La Rochela: memorias de un crimen contra la justicia, 2011, p. 49.
a whiskey, Devia Arias told me he would pay five million pesos for security in Fresno, because he had to campaign there and the guerrilla was present,” a former sub-commander (alias ‘Elkin’) of FOI testified (Verdad Abierta, 2013). Additionally, in the department of Caldas, at least three mayors have been arrested for ties to FOI. Multiple declarations from demobilized paramilitaries suggest that mayors gave cash, markets, petrol and part of the municipality’s contracting jobs (contratación) to the paramilitaries (El Espectador, 2011). While El Gurre partially financed his operations through government contracting, McGuiver willingly choose to stay out of such relations with officials. In a public hearing about the FJLZ’s rent structure, McGuiver responded to the judge’s question whether his front had been involved in contratación in the following way:
“No, no, no… That was a clear choice from my part. I was thought from a very young age that you should not feed corruption – that is, I cannot preach against it [e.g. corruption] and then do the contrary.”24
By avoiding corrupt government officials and contracting deals, McGuiver grew more dependent on locals for resources and support, but was also less likely to be drawn into the government’s counter-insurgency due to La Danta’s relatively peripheral location in Sonsón. For El Gurre this was more difficult as FOI operated in close proximity to a military base in La Dorada, which disallowed him to avoid collusion with security forces in Tolima and Caldas (El Nuevo Día, 2014), a claim that will be elaborated further when the first mechanism linking state control to militia governance emergence, namely participation in government COIN, is discussed.
Similar to Isaza and El Gurre, the commander of the ACMM’s Frente Celestino Mantilla (FCM) was cooperating closely with government officials. Since his arrival to the region in 2001, commander Pájaro had strong ties to police captain Carlos Arturo Téllez – who “gave” twenty-five police grenades to the paramilitaries and ever since was allied to FCM (Verdad Abierta, 2012). Téllez continued selling weapons to the paramilitary front, informed Pájaro about government troop
24 Luis Eduardo Zuluaga (McGuiver) public hearing on 30th of May 2011, video material. “No, no, no… Eso si
fue una de mi parte. Me enseñaron desde muy pequeño que la corupción no se puede alimentar. Osea, yo no puedo predicar claro cuando estoy hacienda lo contrario.”
mobilization, provided coordinates of guerrilla camps, and facilitated the FCM passage through police and military road blocks.25 Moreover, there are allegations of FCM nexuses with
Colombia’s Administrative Department of Security (Departamento Administrativo de Seguridad, DAS), the security service agency (El Tiempo, 2005). In return for information and government support, Pájaro provided the agency access to a cocaine lab (cocina) in San Fransisco, which the DAS could then – delegated by television crews and cameras – raid and terminate.
In the above, I discussed the para-politics many of the ACMM front commanders were involved in. The FJLZ stands out as a paramilitary front willingly avoiding such ties both because it had the capacity to do so, thanks to its geographical isolation, and due to commander McGuiver’s policy to minimize such nexuses. After all, involvement in government corruption delegitimizes his own rule. “The mayors would not come down here because they knew they would have trouble with me,” McGuiver supposedly said.26 While the other fronts had overlapping ties with
government elites and officials, cooperated with government actors, and in the case of the FOI were operating in close proximity to the military 16th Patriotic Battalion (Batallón Patriotas) as well,
the FJLZ maintained militia autonomy. Now I turn to discussing the three mechanisms linking low state control to a higher likelihood of paramilitary governance emergence, and trace why low state control makes governance likely. First, high state control in terms of overlapping elite networks and proximate military base presence draws militias into the government’s counter-insurgency campaign.
Mechanism 1: Participation in government counter-insurgency
Luis Antonio Meneses, former army lieutenant and paramilitary leader in Puerto Boyacá, better known by his alias ‘Ariel Otero’, famously said that the peasant self-defenses were “an anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti- anti-
25 Ibid.
subversive government policy”.27 According to Ariel Otero, the higher echelons of the
Colombian military persuaded its rank-and-file members to disassociate themselves from the army and spread across the country to create new anti-rebel paramilitary groups in the early eighties,28 skewing the composition of paramilitaries operating under high state control towards
relatively more ex-militaries (e.g. Mechanism 2). Indeed, dissatisfaction among high-ranking military officials with President Belisario Betancur’s appeasement strategy towards the FARC, and particularly the participation of the guerrilla’s political party Patriotic Union (Unión Patriótica, UP) in elections a year after Betancur signed a cease-fire deal in 1984 (Suhner, 2002: 171; Delgado, 2006; Vásquez, 2006: 316; Ronderos, 2014: 45), convinced military leaders it was necessary to fight a clandestine war (Molano, 2009: 40). As a result, military officials dusted-off an old playbook: the United States’ counter-insurgency Plan Lazo.29 Plan Lazo was a U.S.
counter-insurgency plan the Colombian military had first implemented in 1962 and called for broad
autodefensa civic action programs coupled with military action to “eliminate the independent
guerrilla republics” (Rempe, 1995a; 1995b).30 By colluding with state forces, self-defense groups
such as Isaza’s Los Escopeteros did not only learn how to defend communities against guerrilla attacks, but also “adopted an offensive attitude”.31
Initially, in the first two years after its mobilization in 1977, Los Escopeteros installed several guarding posts at farms that had often been looted by the guerrilla. In a matter of months, Isaza’s men took hold of Puerto Triunfo, Cocorná, and Sonsón, protecting local businessmen and monitoring populations (Verdad Abierta, 2010). They sustained their efforts with voluntary
27 Luis Antonio Meneses, alias ‘Ariel Otero’, Interrogation session. See: Setencia de Ramón Isaza Arango y
otros, p. 224, f.n. 314.
28 Ibid.
29 Interview #10, 23rd of January 2015, Bogotá; and, Interview #11, 17th of December 2014, Barrancabermeja. 30 Also see Maullin (1971) and Human Rights Watch (1996).
31 Corte Interamericana de Derechos Humanos, Caso 19 comerciantes vs. Colombia, 5th of July 2004, paragraph
transactions from cattle ranchers and loggers they protected.32 Isaza was in the private protection
business – providing security as a club good to those willing to pay, rather than a public good society at large benefited from. In return, Los Escopeteros protected their clients’ livelihoods against cattle rustlers, guerrilleros and thieves.33 From the 1980s onwards, criminal investigation suggests,
the group became more and more a mobile counter-insurgency force pursuing the guerrilla across terrain. In accordance to the Olsonian logic underpinning the argument’s first mechanism, the increased mobility of Los Escopeteros was paired with a shift in the paramilitaries’ strategy toward civilians. Previously paid voluntary donations were partially replaced by arbitrary, coerced payments enforced on farmers, traders, and other people with purchasing power.34 Hence, Isaza’s
offensive strategy had acute implications for civilians living under his rule, who became victims of extortion rather than the beneficiaries of private security.
The increasingly offensive strategy of Los Escopeteros was grounded in government incentives. Not only did Isaza admit that the army had provided him with a hit list of over seventy suspected guerrilla collaborators (Caracol, 2008), illuminating the devolution of offensive violence when state control increases, he also publically accepted responsibility for the massacre in La Rochela in 1989, a crime Isaza’s men committed when fighting under the banners of the ACPB, while colloquially being known as “masetos” – henchmen of MAS, a fighting force operating under the clandestine auspices of the Colombian government and powerful narcotics traffickers (El Espectador, 2014). Evidence shows that both militaries and politicians had a stake and demand in Isaza’s killings of two judges and ten legal investigators in La Rochela (Grupo de Memoria Histórica, 2011). That linkages to the state imply not only offense, but fewer incentives to care for civilian welfare too, is apparent in the following testimony of a victim, remembering the words of a paramilitary: “(…) he [a Maseto] told me that they could kill anyone and nobody would
32 Interview #6. Also see: Tribunal Superior de Bogotá, Sala de Justicia y Paz, Sentencia de Ramón Isaza Arango
y otros, chapter ‘Contexto Histórico Y Socio Político: Autodefensas Campesinas Del Magdalena Medio’, p. 271-272. Published 29th of May 2014.
33 Sentencia de Ramón Isaza Arango y otros, p. 237. 34 Ibid., p. 272.
