Of Rivals and Enemies: Military–Police Relations in Internal Armed Conflict in Mexico and Colombia

A Civil–Military Marketplace of Agents, Expanding on the P-A Framework

The very term “civil–military relations” connotes a dyad, a relationship between two sets of actors: “militaries” and “civilians.” Militaries are unique in their ability to provide state-sanctioned violence, their strict and often anti-democratic hierarchies, and privileged expertise over war-making (Janowitz, 2017). For their part, “civilians” in government are disadvantaged by a knowledge deficit regarding defense policy and security strategies (Cottey et al., 2002). While chief executives and their advisors may have some military background, or some previous defense experience, it is very unlikely that they will have the same level of military training and education as their generals and admirals. For this reason, even militaries under what Samuel Huntington described as “objective civilian control” can use their exclusive knowledge to exploit or mislead the civilians who they answer to (Rahbek-Clemmensen et al., 2012; Huntington, 1981).

Huntington’s “objective control,” originating in his studies of the civil–military relations of the United Kingdom and United States, posited that militaries under civilian political control were sufficiently subordinated to democracy. Huntington advised that civilians should leave their militaries to their own devices so as not to interfere in affairs they were ignorant of. Understandably, many of Huntington’s critics came from a region where politically dissatisfied militaries not only inserted themselves into politics but even took control of governance. Latin American case studies from scholars such as Finer and Nordlinger show that a military left to its own devices can become a fatal threat to democracy (see Finer, 2002; Nordlinger, 1976). For this reason, much of civil–military relations has focused on coups.

However, militaries who accept civilian rule, respect democracy, and pose no political danger to democratic regimes can still attempt to gain the upper hand in their interactions with civilian governments. Moving beyond the coup-fallacy (see Croissant et al., 2010), a growing number of scholars have focused on more mundane, day-to-day interactions between civilian governments and their militaries (Croissant et al., 2010). To study routine interactions between civilian governments and their militaries, Peter Feaver imports the principal–agent framework into civil–military relations (Feaver, 2009). Borrowed from economics, the principal–agent (p-a) framework analyzes relationships where one or more actors (the agents) are subordinate to the other (the principal). However, though the agent is the subservient actor, they have expert knowledge for which the principal relies on them. This means that the agent may exploit their knowledge and either “work,” conducting a task in accordance with the principal’s preferences, or mislead and “shirk,” carrying out a task according to their own preferences.

Feaver convincingly demonstrates that the p-a framework is an excellent tool for analyzing the day-to-day interactions between militaries and their civilian principals. He outlines scenarios where militaries may shirk, such as when presenting civilians with incomplete information to affect policy outcomes. He does, however, acknowledge that the p-a framework is not a perfect fit for civil–military relations. One critical feature of the p-a framework, outside of Feaver’s importation, is that the principal has the option to employ multiple agents (Gupta & Romano, 1998). Feaver acknowledges that no such market place exists in his conception of the p-a framework in civil–military relations:

Of course, there is an anomaly in applying the principal-agent framework to the civil-military setting. There is not really a market of agents; the civilian cannot hire from many different militaries to do its work.

Feaver’s characterization of the civil–military setting is particularly true in the United States, the case upon which his theory and empirics are based. That there is no marketplace of agents is not true in instances where militaries are deployed inward toward domestic threats. In such contexts, militaries are deployed alongside a variety of state agents, namely police forces and state-controlled paramilitary forces. The latter set of actors usually provide a set of security services comparable to military capabilities and are armed with military weapons and vehicles. Once these actors, capable of providing such services to civilian governments, are deployed alongside the military, they become relevant alternatives to civilian principals when making security policy decisions.

This study is part of a growing literature of civil–military relations scholarship examining forces beyond the military. One family of scholarship has used the civil–military dyad as a spectrum, placing forces which straddle the police/military distinction on a civilian-to-military scale. From this scholarship, the terms “intermediate forces” and “hybrid forces” have been used to describe gendarmeries and other paramilitary police forces (Alda, 2016b; Pion-Berlin & Trinkunas, 2011). Others have recently used the term “state forces” to refer to the wide set of government-controlled security forces who exist outside of military hierarchies (De Bruin, 2021). While these terms are useful, I argue that they are tailored to the purposes of the scholars who use them. “Intermediate” and “hybrid” forces are useful for scholars trying to identify the nature of a force, whereas De Bruin’s “state forces” is a great term to utilize for an impressive quantitative study concerning the sheer global variety of non-military forces.

My object of focus, however, is neither the nature nor variety of forces. Instead, my research focus is on the agency and relationships between security agents and their principals. The descriptor “alternative security forces” is employed with a focus on agency and relationships in mind. Furthermore, as my cases are Latin American cases, a region where the military has historically been the security agent of first resort (López-Alves, 2000; Loveman, 1999), I argue that the descriptor “alternative” is more than appropriate in the context of this study. ⁴ However, the term “alternative security forces” is most useful under certain scope conditions, namely where there is a high degree of mission overlap between military and non-military forces linked to strong institutionalized militarization of public security. Where the involvement of the military in internal security and policing matters has been institutionalized over the course of decades, and where legal separations between the missions of the armed forces and other forces are ill-defined, civilians approaching other forces as “alternatives” to the military will be more likely. In contrast, in nations such as Argentina and Uruguay, where there is a strict legal separation between military and police missions, it is far less likely that civilians will be considering other forces as “alternatives” to the military. ⁵

To be a meaningful alternative to the military, ASFs must meet two conditions. First, they must report to the same commander-in-chief as the armed forces to be an alternative that the chief executive can consult and deploy. Second, an ASF must have its own hierarchy separate from that of the armed forces, as otherwise they are merely military co-opted or controlled police forces. The first condition is seemingly straightforward but does exclude several actors from the multi-agent model, namely those who do not report directly to a nation’s commander in chief. For instance, local and subnational police forces report to mayors and governors and are therefore unavailable to serve as an alternative to a nation’s commander in chief (see Sabet, 2012). The same is true of non-state actors, such as vigilante or paramilitary squads, who cannot be consulted by national leaders without severe political cost and are not in the state’s formal hierarchy. ⁶ The second condition excludes more actors from being ASFs if they are merely subsidiaries or specialized branches of the armed forces. It is a mistake, for instance, to call Mexico’s newly created “Guardia Nacional” (National Guard) an “alternative” security force as Mexico’s president placed the force first under de-factor military control (the Guardia’s top commander was an active-duty army general and the majority of its officers came from the military) before attempting to formally and legally place the Guardia Nacional under the control of the military in 2024 (Ivey, 2023). ⁷

With these conditions outlined, there are, globally, a variety of security forces which meet them (De Bruin, 2021). These forces are centralized under the control of the chief executive, and usually are controlled by ministries of the interior, security, or law. Charged with policing missions, investigating, and preventing crimes, these forces have military training and equipment which can be mobilized in the case of war, foreign invasion, or participation in joint UN operations. Continental Europe is, for instance, home to several “intermediate” or “hybrid” forces such as France’s Gendarmerie, Spain’s Guardia Civil, and Italy’s Carabinieri. These are militarized police forces which are subject to the control of the executive branch, acting as police in peace time and military auxiliaries in war time (Lutterbeck, 2004; Vizcaíno, 2018). ⁸

In Latin America, there are few nations with gendarmeries who report to the central government. The sole exception is Argentina’s Gendarmeria Nacional, created in 1938 for rural policing, border operations, and assisting in the colonization of Argentina’s interior (Escolar, 2017). Instead, by both necessity and designed, many police forces have taken on military characteristics to fight internal armed threats. For instance, Colombia’s National Police were created in the 1950s for the express purpose of aiding the military in the civil war “La Violencia.” Chile’s Carabineros are another example, as this centralized and paramilitary police force has historically been the primary state agent used in against insurgencies in the Chilean south. Although the militarized nature of both Colombia’s National Police and Chile’s Carabineros have been criticized as potential contributors to both forces committing human rights abuses during recent nationwide protests, these capabilities make these forces viable and capable alternatives to the armed forces.

Case Selection, Design, and Methodology

Data collection for this article was undertaken according to a qualitative, most similar case design. My initial research agenda was to examine how the relationships governments and militaries changed (if they changed at all) when a third actor with comparable skill sets to those of the military were introduced. When conducting fieldwork, consisting of 6 months split evenly between Bogotá, Colombia and Mexico City, Mexico, I interviewed more than 40 respondents consisting of retired and active-duty police and military personnel, as well as presidential and ministerial advisors. Semi-structured interviews allowed for a more conversational tone and lasted anywhere from 1 to 2 hr. ⁹ Although the general subject of each interview was the interactions between police and military forces, participants were interviewed on their past careers working in or studying security politics. This allowed for firsthand testimony to open the “black box” of security policymaking, with the advantage of access to personnel who witnessed the highest levels of interplay between the leaders of militaries, ASFs, and leaders in civilian governments. I also consulted primary documents from Colombia’s Biblioteca del Congreso, the National Police Museum, and Mexican police observatories such as Causa en Común.

I undertook fieldwork research with a greater focus on finding under which conditions civilian power was increased by the presence of a viable alternative to the military. However, qualitative and exploratory research such as that inherent in political science fieldwork lends itself to observational research which in turn raises new and unexpected questions. Likewise, fieldwork is inherently iterative process which relies on the interplay between inductive and deductive reasoning. Kapiszewski, MacLean, and Read describe the fieldwork process thusly: “As scholars observe and learn in the field, they commonly pivot multiple times, shifting from collecting data to analyzing them; from analysis back to research design; and from research design on to data collection, and on to analysis again” (Kapiszewski et al., 2015, p. 24). As such, my object of focus expanded from exclusively on the effects of a viable alternative to the military on civilian power to the relationships between security agents.

Over the course of my fieldwork, I was struck by the much more hostile relationship between Mexico’s security forces than those in Colombia, particularly because eminent civil–military scholars have observed a heated relationship between Colombia’s National Police and its military (Bruneau, 2004). I found that the “tone and tenor” of military-ASF rivalries, a variable which I had not considered prior to my fieldwork, not only affected security outcomes on the ground but also frustrated the degree to which civilians could effectively manage security policymaking. I bring rivalries to the center of this article to argue the importance of inter-agent relationships in affecting the powers of civilian policymakers and security outcomes on the ground. In the next section, I will elaborate on these causal relationships further.

Colombia and Mexico were selected for a most similar research design, where two cases share many characteristics, but differ on an outcome (see Goertz, 2006, p. 181). In the case of Colombia and Mexico, the two are both Latin American democracies who democratized during ongoing internal armed violence. But while Colombia has successfully reduced (not eliminated) armed threats, Mexico has struggled to make any long-term security gains against internal armed belligerents. Although the reality of political science research is that no two cases are completely identical, Colombia and Mexico share much in common and have been used in most similar designs in other recent scholarship (see Alda, 2016b; Ríos-Figueroa, 2016).

The most apparent, and tragic, similarity between Colombia and Mexico is their status as regional epicenters of the international drug trade and the violence associated with it. In both states, the spill-over violence of the drug trade has blurred the lines of criminal and political conflict (Alda, 2016b). Although the drug cartels of Mexico did not adopt overtly partisan agendas like Colombia’s leftists’ insurgencies, Mexico’s cartels are sophisticated criminal networks with access to military weaponry capable of pushing the state out of its own territory (see Trejo & Ley, 2020). The Mexican Cartels have also proven capable of corrupting and co-opting the Mexican state, an activity which while not “revolutionary” cannot be construed as anything other than “political” (Trejo & Ley, 2020). Similarly, Colombia’s two largest leftist insurgencies, the Revolutionary Armed Forces of Colombia (FARC) and Popular Liberation Army (ELN), have increasingly involved themselves in the drug trade to fund their conflict with the Colombian State. Right wing paramilitaries, such as the United Self-Defenses (AUG), have enriched themselves on the drug trade in their fight against insurgent groups and state agents.

Both states also democratized in the context of unresolved internal conflicts. Colombia’s democracy emerged in the year 1958, when the 10-year civil war La Violencia created space for banditry and insurgency. Although Mexico’s democratic transition began much later, when the one-party dictatorship of the Institutionalized Revolutionary Part (PRI) was broken in 2000, drug violence between the military and criminal cartels had been intensifying since the 1970s. Although both nations have struggled with to protect human rights during internal armed violence, both remain states where democratic elections are free and fair. This has led several observers to describe both Colombia and Mexico as “violent democracies,” another tragic similarity (Whitehead et al., 2010).

In part, because of the historic violent context of their respective democratizations, both nations have likewise struggled to fully separate the missions of their police and military forces. For instance, article 216 of the Colombian Constitution designates Colombian National Police and the Colombian Armed Forces as part of one “Public Force (Fuerza Publica).” ¹⁰ Article 217 states that the military’s primary duties are not only to protect national sovereignty, but to protect “territorial integrity” and “constitutional order” as well. These latter duties legitimize the military’s internal deployment and are not far removed from the police’s designated duties. These police duties are outlined in Article 218, which state that the police work to maintain the “conditions needed for the public exercise of public liberties and rights, to ensure that the inhabitants of Colombia can live in Peace.” The military’s prescribed mission to protect “constitutional order,” along with the constitution’s designation of the police as the forces to defend constitutional rights, enshrines the constitutional entanglements of Colombia’s security forces. Although Colombia’s Constitutional Court has distinguished the police as a “reactive and passive” force and the military as “an active force of attack” (Ríos-Figueroa, 2016, p. 76), military and police missions have remained overlapped since before Colombia’s transition to democracy in 1958.

Likewise, while there has been debate regarding the constitutionality of Mexico’s internal use of the armed forces, the reality is that presidents have disregarded any supposed police-military separation since the end of the Mexican Revolution in the 1920s. While the text of Article 129 limits the military’s “peace time” jurisdiction to issues concerning military discipline, the PRI dictatorship (roughly 1929–2000) used the military internal repression and counter-crime operations (López-González, 2012). ¹¹ The internal use of the military has continued past the end of one-party PRI rule, and Mexico’s supreme court has ruled that the internal deployment of the armed forces does not violate Article 129 (Díez & Nicholls, 2006, p. 39). The separation between “soldier” and “police” in Mexico has only become murkier during the presidency of Andrés Manuel López Obrador (AMLO, 2018–2024), who’s “National Guard” (Guardia Nacional) has been folded into military command. Rather than a clear bifurcation between “police” and “military,” Mexico’s legal framework and politicians have deliberately blurred these lines.

Finally, Colombia and Mexico have both been strategic focuses of the United States during the 21st century. From 2000 to 2017, the United States continually reauthorized strategic aid to both the Colombian Military and Police through “Plan Colombia.” ¹² From 2008 to 2021, the United States likewise participated in the Mérida Initiative with Mexico. ¹³ Both Plan Colombia and the Mérida Initiative were substantial security aid packages and saw the United States work with both nations’ respective military and police forces. Both plans were designed to combat drug trafficking, facilitate cooperation between security forces, and free-up U.S. military equipment and training for knowledge dissemination. While this study does not focus on the role of the United States in either Colombia or Mexico, foreign aid and the interest of the United States have been the focus of analysts who see both Plan Colombia and the Mérida Initiative as drivers of militarization. ¹⁴

However, as similar as Colombia and Mexico are, there are significant differences between the two which should be acknowledged in good faith. The most obvious difference is the longevity of Colombia’s democracy relative to that of Mexico. Colombia’s transitioned to its current democracy in 1958 in a pacted negotiation between an outgoing military regime and an alliance of civilian political parties. This impacted not only the compatibility of case comparison but also the data availability. In Colombia, the testimonies of defense ministers have been filed since the 1950s with Colombia’s congress and are available to researchers (including myself) at the Biblioteca de Congreso in Bogotá. Likewise, the Museo de Policia Nacional provided access to documents on the history of the Colombian National Police and its operations. Although there is no publicly available archive for Mexico’s police, the public think-tank Causa en Común was a valuable resource for documents on the history of Mexico’s Federal Police.

Another important difference is the institutional design of security politics in both countries. The two share in common that their security arrangements are unique in Latin America. Colombia is the only democracy in Latin America whose national police have remained under the portfolio of the Ministry of Defense as opposed to the Ministry of the Interior (Alda, 2016a). For its part, Mexico has been described as “the odd man out,” for two reasons: (1) its army and navy are bifurcated across the Secretariat of Defense and the Secretariat of the Navy, and the Secretaries in charge of these bureaucracies are active-duty generals and admirals (Pion-Berlin, 2009, p. 18). Colombia’s Minister of Defense, who acts for the president in managing both the Colombian National Police and the Military, has been a civilian since 1991, whereas the only civilian who exercises meaningful control over the armed forces in Mexico is the president (Díez, 2008; Díez & Nicholls, 2006; López-González, 2012). Likewise, Colombia’s judiciary has been a more aggressive arbiter of civil–military relations than its counterparts in Mexico (Ríos-Figueroa, 2016). Although Mexico’s Supreme Court has been reactive and limited in exercising powers of civil–military relations, Ríos Figueroa argues, as do I, that Mexico and Colombia are compatible in a most similar case design; as Mexico and Colombia share an internal crisis faced by the military, a dejure commitment to human rights, and competitive democratic elections (Ríos-Figueroa 2016, pp. 156–157).

Finally, though Colombia has had a greater number of civilians participating in security policymaking with the armed forces in Mexico, I will use my case study sections to show that Mexican policymakers have made significant efforts to remedy this deficit. Particularly during the presidency of Felipe Calderón (2006–2012), there was an effort to increase the number of civilian security advisors available to the president and to increase cooperation between these advisors and the armed forces. However, I will show that the zero-sum rivalry between Mexico’s armed forces and Calderón’s commander of the Federal Police undermined this cooperation, and consequently set the stage for future presidents to marginalize their police forces. In Colombia, by contrast, though the position of Minister of Defense transitioned to civilian hands in 1991, it would be inaccurate to say the Ministry was thoroughly civilianized at the moment of transition. I will show, by Minister Rafael Pardo’s own account, that the military itself suggested elevating the Colombian National Police (Pardo, 1996). Here, an interservice rivalry and joint operations between the Police and Military invited greater civilian involvement in security politics.

The Costs and Consequences of Structural Competition

This study focuses on the impact the “tone and tenor” of the rivalries between militaries and ASFs. It builds off the logic of the principal–agent framework and expands the civil–military dyad to a triadic relationship between one principal (civilian governments) and two agents: militaries and ASFs. I begin with the observation that when two agents (in this case militaries and ASFs) provide similar services to the same principal, they are by the very nature of their shared skillset and limited clientele locked into a relationship I describe “structural competition.” This competition is less a consequence of any preexisting individual, institutional, or cultural enmity and instead the consequence of targeting a limited clientele when providing similar goods. Keeping within the logic of the p-a framework, two agents providing similar services will need to make the case to their principal that they are more deserving of the principal’s favor and patronage. However, while this rivalry is usually fruitful for civilians looking to narrow the information gap between themselves and their armed agents, it can worsen security conditions on the ground. Institutional animosity and turf wars can impede civilians’ abilities to gain accurate information relevant to their decision making. Civilians then need to be careful when exploiting this rivalry, being attentive to it while attempting to foster the most amiable relationship possible.

When conceptualizing military-ASF rivalries, I begin by building off preexisting concepts of rivalry and competition within the military itself. Samuel Huntington demonstrated competition between the army and navy helped expand congressional oversight over the United States military (Huntington, 1961), while Feaver includes interservice rivalries as one mechanism through which civilians can expand their oversight powers (Feaver, 2009). Similarly, Harold Trinkunas has shown that Venezuela exploited competitions between the army and navy to dismantle the military’s political power in the wake of that nation’s transition to democracy in 1958 (Trinkunas, 2002). While Trinkunas’ analysis of Venezuela is important in demonstrating that competition can be exploited by civilian governments, Huntington and Feaver’s analyses are important in illustrating that forces in competition are not necessarily beyond cooperation. Huntington and Feaver describe what is commonly known as an “interservice rivalry,” in their works exemplified in the relationship between the United States Army and Navy in which the two compete but are also capable of cooperating with each other through joint operations and intelligence sharing (Feaver, 2009; Huntington, 1961).

Were the army and navy unable to cooperate, joint operations, intelligence collection, and intelligence sharing would become much more difficult, perhaps even impossible. Rivalries in which intelligence and institutional interests are too zealously coveted not only endanger the capacity for jointness but can create critical gaps in intelligence that hinder any positive security outcomes on the ground. We need to look no further than the rivalries which consumed U.S. intelligence agencies prior to the 9/11 terror attacks (Wright, 2006). These rivalries, which privilege institutional interests over national security, are more zero-sum and can, as evidenced, be disastrous for national security. It is then important to not only acknowledge the possibility of enmity between security forces, but its consequences as well. To avoid this, presidents need to be careful in adopting a hands-on management style that does not fatally favor one agent over the other, while also creating incentives for cooperation.

I present two types of rivalries between militaries and ASFs. The first, an interservice rivalry, is a relationship where both militaries and ASFs compete for prestige, resources, and the favor of a commander-in-chief, but share mutual goals which supersede their individual institutional interests. In the case of security agents, the greatest shared interest is national security and respect for civilian control. When national security is at risk, or when civilian commanders so order it, agents in an interservice rivalry are not beyond cooperating with each other, sharing intelligence, and conducting operations at the orders of civilian governments. Doing so will result in better national security outcomes, as forces may not only cooperate, but jointly execute operations which they would otherwise be unable to. Such an arrangement is ideal for civilians, who will gain the benefit of their ability to leverage agents against each other while also allowing them to mandate cooperation when deemed necessary.

The second, a zero-sum rivalry, is far more deleterious and is characterized by suspicion and hostility between agents. Institutional interests are prioritized over shared goals, and cooperation becomes difficult or even impossible when civilians order it. This in turn generates suboptimal security outcomes on the ground, as agents cannot coordinate joint strikes against critical enemy targets. For civilian managers attempting to mitigate and manage this animosity, while simultaneously trying to leverage agents against each other to assert, strife between security agents can generate obstacles to their own knowledge gathering. Agents seeking to undermine each other, rather than prioritize national security, may chronically shirk in providing misleading accounts and information as they attempt to undermine the principal’s trust in their competitor. It is then incumbent on civilians to exploit rivalries responsibly, to not prioritize one agent at the risk of alienating the other, and not to feed inter-agency animosity.

I find that the tone and tenor of the rivalries between militaries and ASFs in Colombia and Mexico have led to strengthened civilian management in Colombia, as well as greater security successes on the ground. The reverse is true in Mexico. Whereas Colombia’s military recognized the importance of the police’s superior intelligence gathering capacities, and petitioned civilians to promote the police in security policy deliberations; the relationship between security agencies in Mexico was and remains hostile, particularly at the highest levels of leadership. Whereas an interservice rivalry in Colombia allowed for both civilians to expand their oversight over security policymaking and for agents to share critical information while coordinating in joint operations, a zero-sum rivalry has perpetuated the military monopoly over Mexican security politics.

Interservice Rivalry in Colombia

The contemporary Colombian National Police (henceforth referred to as “CNP”) originate from the aftermath of the Bogotázo, a 1948 riot in the nation’s capital which kicked off the 10-year civil war of “La Violencia.” When conservative president Mario Ospina Pérez saw police officers participate in anti-conservative rallies, he disbanded all existing police forces in Colombia and replaced them with a new police force under the control of the Colombian Army (Aparicio, 2018, p. 47; Esparza, 2022). From 1948 to 1960, this configuration was held, as police professionalization and direct control of the police remained in the hands of the Colombian Army (Interview with Javier Torres). ¹⁵ In 1960, the CNP were separated from the military when President Alberto Lleras Camargo issued Decree 1705, which moved the police out of military hierarchy (Congreso de Colombia, 1961). In 1965, operational command of the police passed to a graduate of the police academy, formalizing the hierarchical separation of the army and CNP (Policía Nacional, 2003).

Since 1965, the CNP have been formally separated from the military’s hierarchy, meeting a critical condition of qualifying as an ASF. Unlike other nominally “civilian” police, the CNP also participated in counterinsurgent and counter-bandit operations during and after “La Violencia,” with military leadership not only supporting police participation in these operations but lauding them as well (Congreso de Colombia, 1959, 1960). This meant that the police would assume military missions and develop military-like capacities as they were forced to go on the offensive and defensive against insurgent threats. However, the Minister of Defense remained an active-duty general until 1991. Though the CNP and the Army both reported to the Minister of Defense, the ministers prioritized the institutional interests of the army over other security forces. The CNP were then charged with missions which the military did not want (Britto, 2020), which ironically led the police to outpace the army’s intelligence capacities as they were repeatedly called upon to arrest drug traffickers, fight insurgents, and break up urban extortion networks while the army sought to preserve its external orientation.

This relationship of “subordinate autonomy,” where the police had operational autonomy but were forced to take part in missions which the military viewed as undesirable, came to an end in the late 1980s and early 1990s, where the duel threats of the urban insurgency M-19 and the “grand cartels” of Calí and Medellín became the foremost national security threats (Interview with General Henry Medina). ¹⁶ The ascent of M-19 and the rise of Pablo Escobar contributed to Colombia’s violent years, as the internal armed conflict moved from the nation’s rural zones and into the capitol. Peace negotiations with M-19 contributed to a movement for a new constitution, leading to the Ministry of Defense transferring from military to civilian hands in 1991. Prior to 1991, Colombian civil–military relations had been governed by the “Lleras Doctrine,” a pact between military leaders and the civilian government that agreed that neither would interfere in the other’s affairs in return for the military abstaining from governance (Borrero Mansilla, 2019a, 2019b, pp. 30–33; Pardo, 1996). This civil–military pact is named for President Alberto Lleras Camargo, who in a 1958 speech to military personnel said:

I do not want the armed forces to decide how to govern the Nation, in place of what the people decide. But I also do not want, in any manner, for politicians to decide how to manage the armed forces, in their technical function, in their discipline, in their rules, and in their personnel. (Pardo, 1996)

There was then no strong precedent of robust civilian control in 1991, nor was there a strong tradition of separating the missions of police and military. In fact, according to Rafael Pardo, Colombia’s first civilian minister of defense since La Violencia, the decision to name a civilian to the position of minister of defense was not grounded in any new constitutional requirements, but instead in the political priorities of President Cesar Gaviria (Pardo, 1996). This is not an indication of any institutionalized civilian strength, and indeed Pardo, in his memoir Del Primer Mano, describes a process of on-the-job-education that relied on meeting with police and military leaders both in groups and individually. In recognizing the strengths of the police and military in the 1990s, Pardo (1996) explains that he recognized that the police had the best counter-narcotics intelligence operations, while the military had a greater number of forces. In meeting with police and military commanders over of a series of days, Pardo reveals that the idea to elevate the police to co-equals in operations in the fight against the Cartel of Medellín came from the military itself. He writes: “The key idea came from the army: create a joint command between the police and the army, making sure that neither force needed to submit to other” (Pardo, 1996). The now famous “Bloques de Busqueda,” joint police-military task forces, coordinated searches and strikes against the Cartel of Medellín and proved critical in bringing down Pablo Escobar and the cartel of Medellín (Pardo, 1996). That the suggestion to extensively involve the police in the fight against the Cartel of Medellín came from the military may be taken as an indication that military leaders were aware of and respected police capabilities. That the command of the Bloques de Busqueda placed the military and police as co-equals is an even stronger indication of an amiable relationship between the two forces, one which predated the transfer of the ministry of defense to civilian hands in 1991. Pardo, for his part, explains that the inclusion of the police allowed him to become a more informed defense minister (Pardo, 1996).

However, competition and rivalry between CNP and military have remained. Thomas Bruneau, commissioned by the Colombian government to draft a report on the quality of its defense ministry, makes note of serious contentions between the police and military during the presidency of Álvaro Uribe (2002–2010). In interviewing defense officials in the early years of Uribe’s government, Bruneau found that the military resisted cooperating with the police (though he noted the branches of the armed services avoided cooperating with each other even more) and police members felt the military treated them as a “secondary military” (Bruneau, 2004). Other defense officials and experts confirmed that the police and military believe they are in competition for the attention of civilians. Armando Borrero, an advisor to the same Ministers of Defense who Bruneau interviewed, explained his view to this author that the Civilian Ministers of Defense needed a “sub-minister of police.” However, when he pitched the idea in a meeting, the leaders of the armed forces indicated that they “would want one too” (Interview with Armando Borrero). ¹⁷ This indicates that while the relationship between the Colombian Police and Military has been productive, it would be a mischaracterization to describe it as “harmonious.”

The interservice rivalry between Colombia’s military and the CNP, however, has proven productive not only for Colombia’s civilian leadership, but fruitful for rivals pursuing their mutual interests. General Rafael Torres, a leader in Colombia’s Marine Corps who was quite successful in eliminating paramilitary threats, described police intelligence as “essential” in joint operations and verified that “some of the best operations were based on police intelligence” (Interview General Torres). ¹⁸ Data from the police archives indicates that not only were the CNP involved in joint operations against the cartels of Medellín and Calí, but were essential in some of the most critical strikes against the FARC during the governments of Álvaro Uribe and Juan Manuel Santos (2010–2018). Police intelligence located Rual Reyes, the ideological second command of the FARC, in Ecuador in 2008 and “El Mono Joyjoy” in 2010 (Policía Nacional, 2013; Sánchez, 2017). The use of police intelligence was supplemented with military force, allowing a successful operation which proved critical in hollowing out FARC leadership. According to police general Oscar Naranjo, the Commander of the Colombian National Police during the beginning of President Santos’ government, Joyjoy’s death was a blow to the FARC’s morale and was the beginning of a series of kingpin strikes carried out with police intelligence and joint-police military operations (Sánchez, 2017). The result was a kingpin strategy which helped pressure the FARC to the negotiation table (Sánchez, 2017).

A police-military rivalry in Colombia has neither overwhelmed civilian political leadership nor hindered strategic progress on the ground. Indeed, the competition between the two very much resembles an interservice rivalry within the armed forces, and the two forces have been able to combine their capacities to severely reduce the strength of non-state challengers. This is a rivalry which has proven productive for Colombia, although much work remains to be done on transitioning the nation away from its internal armed conflict and into peacetime.

Zero-Sum Rivalries in Mexico

Mexico has struggled with the need to create and maintain alternatives to the military since 2000. Although in public both politicians and military commanders have repeatedly claimed that counter crime missions are undesirable, the military has historically remained the primary armed agent of choice. While there has been no shortage of attempts to create ASFs, namely the Federal Police (2008–2018), military reluctance to cooperate with non-military actors and civilian deference to the military have hindered the progress of any alternative agent. This military reluctance, coupled with presidents who are unwilling to exercise their own decision-making power, is indicative of a zero-sum rivalry between the military and police. This zero-sum rivalry has contributed to limited security progress on the ground, the notable lack of nationwide law enforcement (non-military) capacities, and presidential deference to the military at the expense of the president’s own powers.

Mexico’s history of centrally controlled police forces, who could serve as ASFs, is more limited in comparison to Colombia’s. After the Revolution (1911–1918), a cadre of revolutionary warlords came under the fold of the National Revolution Party (PNR). The PNR would eventually become the Party of the Revolution (PRI) and would govern Mexico for 71 years in a single party dictatorship. Competition between police and military forces was rare, and when militaries did interact with local police forces it was often to replace them at the behest of the central government (Davis, 2006). Control of law enforcement fell to lower-level PRI officials or, in rare cases, members of opposition parties who managed to secure local or state-level office (Davis, 2006). It was common for different police forces to be loyal to distinct and particular partisan principals and to attack their political rivals (Davis, 2006). Consequently, presidents often used the military not necessarily to fight or deter crime, but to subvert and deter potential political rivals.

The final PRI presidents expanded military primacy in counter-crime operations while also setting up the first potential off-ramp from militarization. Carlos Salinas (1988–1994) formalized and expanded the military’s role in anti-narcotics operations, at the behest of the U.S. government (López-González, 2012). Ernesto Zedillo (1994–2000), for his part, recognized the need for a nationwide police force able to respond to the increasing threat of drug trafficking and to serve as an alternative to reliance on the military (Esparza, 2022). This lead to the creation of Mexico’s first ASF, the Federal Preventive Police (PFP), which lasted through the last PRI presidency in 2000 to 2008, when President Felipe Calderón (2006–2012) replaced the PFP with the more simply named Federal Police (MPF).

Tensions between the police and the military in Mexico came almost immediately after the end of the PRI era. Vicente Fox (2000–2006), Mexico’s first non-PRI president, made several steps toward reforming Mexican policing. He removed the PFP from the Ministry of the Interior and placed the force under the control of its own ministry, the Secretariat for Public Security (SSP). While this step may be interpreted as elevating the role of an ASF, military–police rivalries manifested very quickly as Fox made several puzzling leadership decisions. Although Fox nominated Alejandro Gertz, a controversial civilian drug warrior with a history in law enforcement, to head the SSP, the president also nominated an Army General, General Rafael Macedo, to head the attorney general’s office. “You had this weird moment,” Bernardo León, an aid to Gertz in the SSP, “Where Dr. Gertz was in charge of the PFP, a largely militarized force. And General Macedo was in charge of the attorney general’s office, the chief civilian law enforcement officer” (Interview with Bernardo León).

What happened next was evidence of a zero-sum rivalry mindset between the military and PFP. Gertz, on a collision course with the armed forces, sought to expand the powers of the PFP to allow it to receive and investigate criminal complaints, something that only the Attorney General’s Office could do at the time. Under the control of General Macedo, this meant that the military could dominate both external defense and interior security/law enforcement. Fox, shirking his responsibilities as president, said he would allow the reform if all cabinet members granted it unanimous support (Ivey, 2023). León describes the result succinctly: “The three military men [the Secretary of Defense, the Secretary of the Navy, and the Attorney General] objected. So, Gertz left [the administration]” (Interview with Bernardo León). That these men were the only cabinet members to do so indicates a shared interest distinct from the rest of the cabinet. Being that their only shared attribute were being military officers indicates a common cause and interest not shared with other cabinet members. The result was the military sabotage of police reform and the continuation of military dominance of internal security through the personage of Attorney General Macedo.

Vicente Fox’s successor, Felipe Calderón (2006–2012), had more success in bolstering alternatives to the armed forces. Calderón has been criticized for deploying the military to counter narcotics traffickers at the outset of his presidency, with some scholars suggesting that this was little more than an attempt to bolster his own legitimacy. However, his creation of the Federal Police (MPF) in 2008 was designed explicitly with the intention of replacing the armed forces in the street. This is one reason, if not the biggest reason, the military viewed the Federal Police with deep suspicion. Whereas the PFP only had a force of 8,000 by the end of the Fox Presidency, the MPF had a force of 40,000 by the end of Calderón’s term (Esparza, 2022). Along with strengthening the number of non-military personnel on the streets, Calderón’s security cabinet meetings were held weekly, and included several prominent civilian advisors, including the president’s technical security secretary who often helped the president manage these meetings (interview Calderón Security Advisor 1). ¹⁹

But despite bolstering the strength of the Federal Police and elevating new civilians who the military would have to cooperate with, Calderón’s management style did not ameliorate the zero-sum rivalry between his security agents. This lies in Calderón’s choice for leader of the MPF: his Secretary of Public Security, Genaro García Luna. Although García Luna is now convicted of drug trafficking in the United States, Genaro García Luna was introduced to Calderón as a career law enforcement officer with a sterling reputation. He was the former head of the AFI, a police force styled after the FBI and under the control of General Macedo during his time as attorney general during the Fox presidency. Although on paper an excellent choice for the position, advisors to Calderón and military officials both describe a man who was competitive and off putting. He was openly dismissive of the reform efforts of Calderón’s other civilian appointees, sabotaging their own efforts to centralize security policymaking in his hands (Lessing, 2017). He also clashed with military leadership, who were far from accustomed to an aggressive civilian security expert who had a president’s backing. Individuals present in security cabinet meetings describe tense conversation and an overly aggressive head of public security who was not afraid to butt heads with the secretaries of defense and the navy. According to Guillermo Valdés, then head of Mexico’s top intelligence agency, García Luna’s reputation and personality preceded him. “Genaro even wanted to take control of CISEN,” Valdés explains. “Calderón had to intervene and tell him: ‘Genaro, no’” (Interview with Guillermo Valdés). ²⁰

García Luna’s conflict with other members of the security cabinet forced Calderón to mandate that all security and intelligence agencies would create one unified intelligence report, a move which strengthened civilian oversight of the armed forces by forcing the military to work with other civilians (Interview with Guillermo Valdés). In this way, competition between security agents did indeed strengthen Calderón’s hand. However, even at this point, agencies were deeply distrustful of each other and shirked in sharing their intelligence, coveting their own operations, and in the opinion of at least one scholar, prioritizing the secrecy of their intelligence over national security (Lessing, 2017, p. 227). Whereas in Colombia intelligence sharing between the CNP and the military resulted in successful operations, cooperation between the MPF and the Armed Forces remained minimal in Mexico. Military patrols would occasionally escort Federal Police officers to make arrests. While such military escorts allowed the Federal Police to make arrests and separated the legal (judicial) and military arms of the state, this cooperation was superficial and surface level.

Although Calderón’s successor, Enrique Peña Nieto (2012–2018), made some efforts toward strengthening the capacities of the Federal Police, his administration also sought to differentiate itself from Calderón’s. By the end of the Calderón presidency, García Luna’s reputation and suspected corruption (later validated by a U.S.-led investigation) cast a shadow over the Federal Police. Peña eliminated the Secretariat of Public Security and placed the Federal Police back into the interior ministry. His administration also created the position of “Sub Commissioner of Public Security,” effectively creating two civilian appointees who functioned as filters between the Federal Police and their commander in chief. The army and navy, meanwhile, continued to hold their cabinet positions. This strengthened military power in the government, as greater access to the president meant a greater degree of military influence at the expense of the Federal Police. According to Renato Sales, the last Commissioner of Public Security in the Peña Presidency, Peña removed the MPF from the cabinet at least in part because of both the public reputation of Genaro García Luna and because the military felt that Calderón favored the MPF over the military (interview with Renato Sales). ²¹ In Sales’ words, the president’s move to marginalize the Federal Police restored an “equilibrium” between his agents.

For Manelich Castilla, commissioner of the federal police from 2016 to 2018, the military’s privileged access to the president was a particularly frustrating obstacle in keeping the president informed. In an interview, Castilla offered a comparison to academia to explain how the federal police struggled to keep the president informed:

The presidents do not know what’s going on. Compare it to a university. If you have a dean who is only listening to senior faculty, then decisions will continue to go the same direction, regardless of what junior faculty are doing. (Interview Manelich Castilla) ²²

The result was that the president would often be informed about what was happening within the Federal Police by his military advisors. This institutional marginalization set the stage for Peña’s successor, Andrés Manuel López Obrador (AMLO; 2018–present) to eliminate the Federal Police entirely, replacing it with a “National Guard” under the direct stewardship of the military.

Developments between the Calderón and Peña sexenio do pose a question about the importance of the structure of Mexico’s security institutions and the agency of the president as well as leaders of the Federal Police and armed forces. Certainly, being removed from the presidential cabinet and placed under the head of a sub commissioner of public security limited the agency of the Federal Police, who struggled to inform the president of their own reforms. While there is the possibility that, perhaps, there were leaders of Mexico’s Army and Navy who would have put the legacy of Genaro García Luna aside, and perhaps could have used their agency to advocate for their rival’s inclusion at the highest levels of security policymaking. That is, after all, what happened in Colombia according to Minister of Defense Rafael Pardo, when the Colombian military advocated for the inclusion and elevation of the Colombian National Police.

It is, however, difficult to disentangle agency and structure in Mexico. As far back as the Fox sexenio, military leaders demonstrated hostility toward the idea of empowering a rival law enforcement agency. Although there are certainly military officers who would police reform over the militarization of law enforcement, those at the highest levels of leadership have repeatedly taken actions which have entrenched and widened military control over internal security. Certainly, the appointment of such abrasive of a person as Genaro García Luna gave military leadership little reason to change their minds regarding the police. And there exists the possibility that if a more conciliatory, less abrasive Secretary of Public Security were selected, the backlash against the Federal Police would have been less intense. Agency then, by all indications, precedes the structural change which occurred in Mexico to the detriment of the Federal Police.

Analysis and Conclusion

Both Colombia and Mexico have been challenged by internal armed actors. In both countries, policymakers have been called upon to deploy both the armed forces and ASFs against these threats. These agents, reporting to a shared civilian principal in the president, have been deployed alongside each other, although in the case of Mexico their leadership has worked against to each other. While soldiers and their rivals have managed to both compete and cooperate in Colombia, using their skillsets (police intelligence and military capacities) to coordinate joint strikes, a culture of bitter, and zero-sum rivalry has contributed a long-lasting law enforcement gap in Mexico. In this gap, non-state actors have thrived, despite the continual deployment of the armed forces against them.

Colombia and Mexico both show that proactive civilian leadership is critical in ameliorating rivalries and in coordinating inter-agent operations. Although Felipé Calderón created the most robust ASF Mexico had ever seen, and should be credited with this, his choice of someone as combative as Genaro García Luna to lead this force did much to feed a culture of zero-sum rivalry. By openly seeking to dominate security politics, Genaro García Luna gave the military reason to suspect that they would be replaced and to view any potential ASF as a threat to their institutional interests. Although Calderón mandated that the military cooperate with other civilians on his security cabinet, and even reduced military autonomy by mandating a joint intelligence briefing, his support of García Luna embittered military leaders to the Federal Police. This translated toward a grudge toward the Federal Police which incentivized subsequent presidents to first diminish Federal Police access to the president (under Peña) and then remove the force entirely (under AMLO).

Civilian leaders need to note that the individual leadership of top commanders will be especially important in managing rivalries. They should do their best to select ASF and military commanders who have a history of working on joint task forces and demonstrating a commitment to cooperating for the sake of national security. This will not eliminate rivalries, but as long as joint operations are possible, these rivalries will not be harmful to security outcomes on the ground. While it is unclear that Calderón could have done this, lessons from his presidency should advise future policymakers to go out of their way to seek commanders with either a history in or an expressed desire for cooperation with potential rival state agencies.

In contrast, Colombian civilians benefited from the fact that the military never viewed the police as a threat. While there is some irony that the military’s reluctance to conduct anti-crime operations in the 1970s contributed to the police surpassing their intelligence capacities, what is more important is the military admitted in 1991 that the police would be essential in combatting Colombia’s greatest security threats. Furthermore, civilian involvement in coordinating joint operations has proven critical at expanding Colombia’s civilianization of defense. Bruneau, revisiting Colombia in a later study, noted that part of the reason that the “civilianization” of Colombia’s defense ministry occurred was so that civilians could drive the police and military toward a doctrine of “jointness” (Bruneau & Goetze, 2021). A desire for jointness, made possible through an interservice rivalry, has at least partially driven Colombia’s continued “civilianization” which began in 1991.

This study has focused on two democracies facing internal armed threats, the logic being that the mission overlap between militaries and ASFs will be highest there. However, a growing number of democracies are using their armed forces for internal operations even as threats are comparable low. This includes not only Latin American democracies such as Chile and Argentina, both of whom are using their armed forces to intercept drug shipments despite having capable and proficient ASFs, but Western European democracies such as France and Italy as well. Future scholarship will integrate new cases of ASF-military rivalry in peacetime democracies and test how threat level intensifies or deescalates institutional rivalries.